[Chore] Update type annotation of `input_ids` in model forward (#33063)

docs/contributing/model/basic.md

+1 -1

@@ -71,7 +71,7 @@ class MyModel(nn.Module):

71 ```python

72 def forward(

73 self,

74 - input_ids: torch.Tensor,

74 + input_ids: torch.Tensor | None,

75 positions: torch.Tensor,

76 intermediate_tensors: IntermediateTensors | None = None,

77 inputs_embeds: torch.Tensor | None = None,

tests/plugins/vllm_add_dummy_model/vllm_add_dummy_model/my_gemma_embedding.py

+1 -1

@@ -36,7 +36,7 @@ class MyGemma2Embedding(nn.Module):

36

37 def forward(

38 self,

39 - input_ids: torch.Tensor,

39 + input_ids: torch.Tensor | None,

40 positions: torch.Tensor,

41 intermediate_tensors: IntermediateTensors | None = None,

42 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/afmoe.py

+2 -2

@@ -425,7 +425,7 @@ class AfmoeModel(nn.Module):

425

426 def forward(

427 self,

428 - input_ids: torch.Tensor,

428 + input_ids: torch.Tensor | None,

429 positions: torch.Tensor,

430 intermediate_tensors: IntermediateTensors | None = None,

431 inputs_embeds: torch.Tensor | None = None,

@@ -675,7 +675,7 @@ class AfmoeForCausalLM(nn.Module, SupportsPP, SupportsLoRA):

675

676 def forward(

677 self,

678 - input_ids: torch.Tensor,

678 + input_ids: torch.Tensor | None,

679 positions: torch.Tensor,

680 intermediate_tensors: IntermediateTensors | None = None,

681 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/apertus.py

+1 -1

@@ -542,7 +542,7 @@ class ApertusForCausalLM(nn.Module, SupportsLoRA, SupportsPP):

542

543 def forward(

544 self,

545 - input_ids: torch.Tensor,

545 + input_ids: torch.Tensor | None,

546 positions: torch.Tensor,

547 intermediate_tensors: IntermediateTensors | None = None,

548 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/arcee.py

+1 -1

@@ -394,7 +394,7 @@ class ArceeForCausalLM(nn.Module, SupportsLoRA, SupportsPP):

394

395 def forward(

396 self,

397 - input_ids: torch.Tensor,

397 + input_ids: torch.Tensor | None,

398 positions: torch.Tensor,

399 intermediate_tensors: IntermediateTensors | None = None,

400 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/arctic.py

+2 -2

@@ -406,7 +406,7 @@ class ArcticModel(nn.Module):

406

407 def forward(

408 self,

409 - input_ids: torch.Tensor,

409 + input_ids: torch.Tensor | None,

410 positions: torch.Tensor,

411 intermediate_tensors: IntermediateTensors | None,

412 inputs_embeds: torch.Tensor | None = None,

@@ -460,7 +460,7 @@ class ArcticForCausalLM(nn.Module, SupportsPP, SupportsQuant):

460

461 def forward(

462 self,

463 - input_ids: torch.Tensor,

463 + input_ids: torch.Tensor | None,

464 positions: torch.Tensor,

465 intermediate_tensors: IntermediateTensors | None = None,

466 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/aria.py

+1 -1

@@ -629,7 +629,7 @@ class AriaForConditionalGeneration(nn.Module, SupportsMultiModal):

629

630 def forward(

631 self,

632 - input_ids: torch.Tensor,

632 + input_ids: torch.Tensor | None,

633 positions: torch.Tensor,

634 intermediate_tensors: IntermediateTensors | None = None,

635 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/audioflamingo3.py

+1 -1

@@ -609,7 +609,7 @@ class AudioFlamingo3ForConditionalGeneration(

609

610 def forward(

611 self,

612 - input_ids: torch.Tensor,

612 + input_ids: torch.Tensor | None,

613 positions: torch.Tensor,

614 intermediate_tensors: IntermediateTensors | None = None,

615 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/aya_vision.py

+1 -1

@@ -420,7 +420,7 @@ class AyaVisionForConditionalGeneration(nn.Module, SupportsMultiModal, SupportsP

420

421 def forward(

422 self,

423 - input_ids: torch.Tensor,

423 + input_ids: torch.Tensor | None,

424 positions: torch.Tensor,

425 intermediate_tensors: IntermediateTensors | None = None,

426 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/bagel.py

+1 -1

@@ -507,7 +507,7 @@ class BagelForConditionalGeneration(

507

508 def forward(

509 self,

510 - input_ids: torch.Tensor,

510 + input_ids: torch.Tensor | None,

511 positions: torch.Tensor,

512 intermediate_tensors: IntermediateTensors | None = None,

513 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/baichuan.py

+2 -2

@@ -311,7 +311,7 @@ class BaiChuanModel(nn.Module):

311

312 def forward(

313 self,

314 - input_ids: torch.Tensor,

314 + input_ids: torch.Tensor | None,

315 positions: torch.Tensor,

316 intermediate_tensors: IntermediateTensors | None,

317 inputs_embeds: torch.Tensor | None = None,

@@ -428,7 +428,7 @@ class BaiChuanBaseForCausalLM(nn.Module, SupportsLoRA, SupportsPP, SupportsQuant

428

429 def forward(

430 self,

431 - input_ids: torch.Tensor,

431 + input_ids: torch.Tensor | None,

432 positions: torch.Tensor,

433 intermediate_tensors: IntermediateTensors | None = None,

434 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/bailing_moe.py

+2 -2

@@ -440,7 +440,7 @@ class BailingMoeModel(nn.Module):

440

441 def forward(

442 self,

443 - input_ids: torch.Tensor,

443 + input_ids: torch.Tensor | None,

444 position_ids: torch.Tensor,

445 intermediate_tensors: IntermediateTensors | None,

446 inputs_embeds: torch.Tensor | None = None,

@@ -611,7 +611,7 @@ class BailingMoeForCausalLM(nn.Module, SupportsPP, SupportsLoRA):

611

612 def forward(

613 self,

614 - input_ids: torch.Tensor,

614 + input_ids: torch.Tensor | None,

615 positions: torch.Tensor,

616 intermediate_tensors: IntermediateTensors | None = None,

617 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/bamba.py

+2 -2

@@ -311,7 +311,7 @@ class BambaModel(nn.Module):

311

312 def forward(

313 self,

314 - input_ids: torch.Tensor,

314 + input_ids: torch.Tensor | None,

315 positions: torch.Tensor,

316 intermediate_tensors: IntermediateTensors | None = None,

317 inputs_embeds: torch.Tensor | None = None,

@@ -493,7 +493,7 @@ class BambaForCausalLM(

493

494 def forward(

495 self,

496 - input_ids: torch.Tensor,

496 + input_ids: torch.Tensor | None,

497 positions: torch.Tensor,

498 intermediate_tensors: IntermediateTensors | None = None,

499 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/bert_with_rope.py

+1 -1

@@ -475,7 +475,7 @@ class BertWithRope(nn.Module, SupportsQuant):

475

476 def forward(

477 self,

478 - input_ids: torch.Tensor,

478 + input_ids: torch.Tensor | None,

479 positions: torch.Tensor,

480 intermediate_tensors: IntermediateTensors | None = None,

481 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/blip2.py

+1 -1

@@ -641,7 +641,7 @@ class Blip2ForConditionalGeneration(

641

642 def forward(

643 self,

644 - input_ids: torch.Tensor,

644 + input_ids: torch.Tensor | None,

645 positions: torch.Tensor,

646 intermediate_tensors: IntermediateTensors | None = None,

647 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/bloom.py

+2 -2

@@ -276,7 +276,7 @@ class BloomModel(nn.Module):

276

277 def forward(

278 self,

279 - input_ids: torch.Tensor,

279 + input_ids: torch.Tensor | None,

280 position_ids: torch.Tensor,

281 intermediate_tensors: IntermediateTensors | None,

282 inputs_embeds: torch.Tensor | None = None,

@@ -358,7 +358,7 @@ class BloomForCausalLM(nn.Module, SupportsPP, SupportsQuant):

358

359 def forward(

360 self,

361 - input_ids: torch.Tensor,

361 + input_ids: torch.Tensor | None,

362 positions: torch.Tensor,

363 intermediate_tensors: IntermediateTensors | None = None,

364 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/chameleon.py

+1 -1

@@ -994,7 +994,7 @@ class ChameleonForConditionalGeneration(

994

995 def forward(

996 self,

997 - input_ids: torch.Tensor,

997 + input_ids: torch.Tensor | None,

998 positions: torch.Tensor,

999 intermediate_tensors: IntermediateTensors | None = None,

1000 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/chatglm.py

+2 -2

@@ -362,7 +362,7 @@ class ChatGLMModel(nn.Module, SupportsQuant):

362

363 def forward(

364 self,

365 - input_ids: torch.Tensor,

365 + input_ids: torch.Tensor | None,

366 positions: torch.Tensor,

367 intermediate_tensors: IntermediateTensors | None = None,

368 inputs_embeds: torch.Tensor | None = None,

@@ -491,7 +491,7 @@ class ChatGLMForCausalLM(ChatGLMBaseModel, SupportsLoRA, SupportsPP, SupportsQua

491

492 def forward(

493 self,

494 - input_ids: torch.Tensor,

494 + input_ids: torch.Tensor | None,

495 positions: torch.Tensor,

496 intermediate_tensors: IntermediateTensors | None = None,

497 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/cohere2_vision.py

+1 -1

@@ -446,7 +446,7 @@ class Cohere2VisionForConditionalGeneration(nn.Module, SupportsMultiModal, Suppo

446

447 def forward(

448 self,

449 - input_ids: torch.Tensor,

449 + input_ids: torch.Tensor | None,

450 positions: torch.Tensor,

451 intermediate_tensors: IntermediateTensors | None = None,

452 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/commandr.py

+2 -2

@@ -312,7 +312,7 @@ class CohereModel(nn.Module):

312

313 def forward(

314 self,

315 - input_ids: torch.Tensor,

315 + input_ids: torch.Tensor | None,

316 positions: torch.Tensor,

317 intermediate_tensors: IntermediateTensors | None,

318 inputs_embeds: torch.Tensor | None = None,

@@ -438,7 +438,7 @@ class CohereForCausalLM(nn.Module, SupportsLoRA, SupportsPP, SupportsQuant):

438 @torch.no_grad()

439 def forward(

440 self,

441 - input_ids: torch.Tensor,

441 + input_ids: torch.Tensor | None,

442 positions: torch.Tensor,

443 intermediate_tensors: IntermediateTensors | None = None,

444 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/dbrx.py

+2 -2

@@ -361,7 +361,7 @@ class DbrxModel(nn.Module):

361

362 def forward(

363 self,

364 - input_ids: torch.Tensor,

364 + input_ids: torch.Tensor | None,

365 position_ids: torch.Tensor,

366 intermediate_tensors: IntermediateTensors | None,

367 inputs_embeds: torch.Tensor | None = None,

@@ -462,7 +462,7 @@ class DbrxForCausalLM(nn.Module, SupportsPP):

462

463 def forward(

464 self,

465 - input_ids: torch.Tensor,

465 + input_ids: torch.Tensor | None,

466 positions: torch.Tensor,

467 intermediate_tensors: IntermediateTensors | None = None,

468 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/deepseek_mtp.py

+1 -1

@@ -213,7 +213,7 @@ class DeepSeekMTP(nn.Module, DeepseekV2MixtureOfExperts):

213

214 def forward(

215 self,

216 - input_ids: torch.Tensor,

216 + input_ids: torch.Tensor | None,

217 positions: torch.Tensor,

218 hidden_states: torch.Tensor,

219 intermediate_tensors: IntermediateTensors | None = None,

vllm/model_executor/models/deepseek_ocr.py

+1 -1

@@ -562,7 +562,7 @@ class DeepseekOCRForCausalLM(nn.Module, SupportsMultiModal, SupportsPP, Supports

562

563 def forward(

564 self,

565 - input_ids: torch.Tensor,

565 + input_ids: torch.Tensor | None,

566 positions: torch.Tensor,

567 intermediate_tensors: IntermediateTensors | None = None,

568 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/deepseek_v2.py

+2 -2

@@ -1085,7 +1085,7 @@ class DeepseekV2Model(nn.Module):

1085

1086 def forward(

1087 self,

1088 - input_ids: torch.Tensor,

1088 + input_ids: torch.Tensor | None,

1089 positions: torch.Tensor,

1090 intermediate_tensors: IntermediateTensors | None,

1091 inputs_embeds: torch.Tensor | None = None,

@@ -1255,7 +1255,7 @@ class DeepseekV2ForCausalLM(

1255

1256 def forward(

1257 self,

1258 - input_ids: torch.Tensor,

1258 + input_ids: torch.Tensor | None,

1259 positions: torch.Tensor,

1260 intermediate_tensors: IntermediateTensors | None = None,

1261 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/deepseek_vl2.py

+1 -1

@@ -614,7 +614,7 @@ class DeepseekVLV2ForCausalLM(nn.Module, SupportsMultiModal, SupportsPP):

614

615 def forward(

616 self,

617 - input_ids: torch.Tensor,

617 + input_ids: torch.Tensor | None,

618 positions: torch.Tensor,

619 intermediate_tensors: IntermediateTensors | None = None,

620 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/dots1.py

+2 -2

@@ -394,7 +394,7 @@ class Dots1Model(nn.Module):

394

395 def forward(

396 self,

397 - input_ids: torch.Tensor,

397 + input_ids: torch.Tensor | None,

398 positions: torch.Tensor,

399 intermediate_tensors: IntermediateTensors | None,

400 inputs_embeds: torch.Tensor | None = None,

@@ -538,7 +538,7 @@ class Dots1ForCausalLM(nn.Module, SupportsPP, SupportsLoRA):

538

539 def forward(

540 self,

541 - input_ids: torch.Tensor,

541 + input_ids: torch.Tensor | None,

542 positions: torch.Tensor,

543 intermediate_tensors: IntermediateTensors | None = None,

544 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/dots_ocr.py

+1 -1

@@ -754,7 +754,7 @@ class DotsOCRForCausalLM(nn.Module, SupportsMultiModal, SupportsPP, SupportsLoRA

754

755 def forward(

756 self,

757 - input_ids: torch.Tensor,

757 + input_ids: torch.Tensor | None,

758 positions: torch.Tensor,

759 intermediate_tensors: IntermediateTensors | None = None,

760 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/eagle2_5_vl.py

+1 -2

@@ -432,7 +432,7 @@ class Eagle2_5_VLForConditionalGeneration(

432

433 def forward(

434 self,

435 - input_ids: torch.Tensor,

435 + input_ids: torch.Tensor | None,

436 positions: torch.Tensor,

437 intermediate_tensors: IntermediateTensors | None = None,

438 inputs_embeds: torch.Tensor | None = None,

@@ -440,7 +440,6 @@ class Eagle2_5_VLForConditionalGeneration(

440 ) -> IntermediateTensors:

441 """Forward pass through the model."""

442 if intermediate_tensors is not None:

443 - input_ids = None

443 inputs_embeds = None

444

445 forward_kwargs = {

vllm/model_executor/models/ernie45_moe.py

+2 -2

@@ -466,7 +466,7 @@ class Ernie4_5_MoeModel(nn.Module):

466

467 def forward(

468 self,

469 - input_ids: torch.Tensor,

469 + input_ids: torch.Tensor | None,

470 positions: torch.Tensor,

471 intermediate_tensors: IntermediateTensors | None = None,

472 inputs_embeds: torch.Tensor | None = None,

@@ -728,7 +728,7 @@ class Ernie4_5_MoeForCausalLM(nn.Module, SupportsPP, SupportsLoRA, MixtureOfExpe

728

729 def forward(

730 self,

731 - input_ids: torch.Tensor,

731 + input_ids: torch.Tensor | None,

732 positions: torch.Tensor,

733 intermediate_tensors: IntermediateTensors | None = None,

734 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/ernie45_vl.py

+1 -1

@@ -1650,7 +1650,7 @@ class Ernie4_5_VLMoeForConditionalGeneration(

1650

1651 def forward(

1652 self,

1653 - input_ids: torch.Tensor,

1653 + input_ids: torch.Tensor | None,

1654 positions: torch.Tensor,

1655 intermediate_tensors: IntermediateTensors | None = None,

1656 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/ernie45_vl_moe.py

+2 -2

@@ -565,7 +565,7 @@ class Ernie4_5_VLMoeModel(nn.Module):

565

566 def forward(

567 self,

568 - input_ids: torch.Tensor,

568 + input_ids: torch.Tensor | None,

569 positions: torch.Tensor,

570 intermediate_tensors: IntermediateTensors | None = None,

571 inputs_embeds: torch.Tensor | None = None,

@@ -646,7 +646,7 @@ class Ernie4_5_VLMoeForCausalLM(nn.Module, SupportsPP):

646

647 def forward(

648 self,

649 - input_ids: torch.Tensor,

649 + input_ids: torch.Tensor | None,

650 positions: torch.Tensor,

651 intermediate_tensors: IntermediateTensors | None = None,

652 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/ernie_mtp.py

+1 -1

@@ -164,7 +164,7 @@ class ErnieMTP(nn.Module):

164

165 def forward(

166 self,

167 - input_ids: torch.Tensor,

167 + input_ids: torch.Tensor | None,

168 positions: torch.Tensor,

169 hidden_states: torch.Tensor,

170 intermediate_tensors: IntermediateTensors | None = None,

vllm/model_executor/models/exaone.py

+1 -1

@@ -496,7 +496,7 @@ class ExaoneForCausalLM(nn.Module, SupportsLoRA, SupportsPP):

496

497 def forward(

498 self,

499 - input_ids: torch.Tensor,

499 + input_ids: torch.Tensor | None,

500 positions: torch.Tensor,

501 intermediate_tensors: IntermediateTensors | None = None,

502 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/exaone4.py

+1 -1

@@ -490,7 +490,7 @@ class Exaone4ForCausalLM(nn.Module, SupportsLoRA, SupportsPP):

490

491 def forward(

492 self,

493 - input_ids: torch.Tensor,

493 + input_ids: torch.Tensor | None,

494 positions: torch.Tensor,

495 intermediate_tensors: IntermediateTensors | None = None,

496 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/exaone_moe.py

+1 -1

@@ -549,7 +549,7 @@ class ExaoneMoeForCausalLM(nn.Module, SupportsLoRA, SupportsPP):

549

550 def forward(

551 self,

552 - input_ids: torch.Tensor,

552 + input_ids: torch.Tensor | None,

553 positions: torch.Tensor,

554 intermediate_tensors: IntermediateTensors | None = None,

555 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/falcon.py

+1 -1

@@ -402,7 +402,7 @@ class FalconModel(nn.Module):

402

403 def forward(

404 self,

405 - input_ids: torch.Tensor,

405 + input_ids: torch.Tensor | None,

406 positions: torch.Tensor,

407 intermediate_tensors: IntermediateTensors | None,

408 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/falcon_h1.py

+2 -2

@@ -459,7 +459,7 @@ class FalconH1Model(nn.Module):

459

460 def forward(

461 self,

462 - input_ids: torch.Tensor,

462 + input_ids: torch.Tensor | None,

463 positions: torch.Tensor,

464 intermediate_tensors: IntermediateTensors | None = None,

465 inputs_embeds: torch.Tensor | None = None,

@@ -602,7 +602,7 @@ class FalconH1ForCausalLM(

602

603 def forward(

604 self,

605 - input_ids: torch.Tensor,

605 + input_ids: torch.Tensor | None,

606 positions: torch.Tensor,

607 intermediate_tensors: IntermediateTensors | None = None,

608 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/fuyu.py

+1 -1

@@ -340,7 +340,7 @@ class FuyuForCausalLM(nn.Module, SupportsMultiModal, SupportsPP):

340

341 def forward(

342 self,

343 - input_ids: torch.Tensor,

343 + input_ids: torch.Tensor | None,

344 positions: torch.Tensor,

345 intermediate_tensors: IntermediateTensors | None = None,

346 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/gemma.py

+2 -2

@@ -297,7 +297,7 @@ class GemmaModel(nn.Module):

297

298 def forward(

299 self,

300 - input_ids: torch.Tensor,

300 + input_ids: torch.Tensor | None,

301 positions: torch.Tensor,

302 intermediate_tensors: IntermediateTensors | None,

303 inputs_embeds: torch.Tensor | None = None,

@@ -400,7 +400,7 @@ class GemmaForCausalLM(nn.Module, SupportsLoRA, SupportsPP):

400

401 def forward(

402 self,

403 - input_ids: torch.Tensor,

403 + input_ids: torch.Tensor | None,

404 positions: torch.Tensor,

405 intermediate_tensors: IntermediateTensors | None = None,

406 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/gemma2.py

+1 -1

@@ -410,7 +410,7 @@ class Gemma2ForCausalLM(nn.Module, SupportsLoRA, SupportsPP):

410

411 def forward(

412 self,

413 - input_ids: torch.Tensor,

413 + input_ids: torch.Tensor | None,

414 positions: torch.Tensor,

415 intermediate_tensors: IntermediateTensors | None = None,

416 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/gemma3.py

+1 -1

@@ -494,7 +494,7 @@ class Gemma3ForCausalLM(nn.Module, SupportsLoRA, SupportsPP):

494

495 def forward(

496 self,

497 - input_ids: torch.Tensor,

497 + input_ids: torch.Tensor | None,

498 positions: torch.Tensor,

499 intermediate_tensors: IntermediateTensors | None = None,

500 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/gemma3_mm.py

+1 -1

@@ -618,7 +618,7 @@ class Gemma3ForConditionalGeneration(

618

619 def forward(

620 self,

621 - input_ids: torch.Tensor,

621 + input_ids: torch.Tensor | None,

622 positions: torch.Tensor,

623 intermediate_tensors: IntermediateTensors | None = None,

624 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/gemma3n.py

+4 -4

@@ -704,7 +704,7 @@ class Gemma3nSelfDecoder(nn.Module):

704

705 def forward(

706 self,

707 - input_ids: torch.Tensor,

707 + input_ids: torch.Tensor | None,

708 positions: torch.Tensor,

709 inputs_embeds: torch.Tensor | None = None,

710 per_layer_inputs: torch.Tensor | None = None,

@@ -887,7 +887,7 @@ class Gemma3nTextModel(nn.Module, SupportsQuant):

887

888 def fast_prefill_forward(

889 self,

890 - input_ids: torch.Tensor,

890 + input_ids: torch.Tensor | None,

891 positions: torch.Tensor,

892 inputs_embeds: torch.Tensor | None = None,

893 per_layer_inputs: torch.Tensor | None = None,

@@ -964,7 +964,7 @@ class Gemma3nTextModel(nn.Module, SupportsQuant):

964

965 def normal_forward(

966 self,

967 - input_ids: torch.Tensor,

967 + input_ids: torch.Tensor | None,

968 positions: torch.Tensor,

969 inputs_embeds: torch.Tensor | None = None,

970 per_layer_inputs: torch.Tensor | None = None,

@@ -1131,7 +1131,7 @@ class Gemma3nForCausalLM(nn.Module):

1131

1132 def forward(

1133 self,

1134 - input_ids: torch.Tensor,

1134 + input_ids: torch.Tensor | None,

1135 positions: torch.Tensor,

1136 *,

1137 per_layer_inputs: torch.Tensor | None = None,

vllm/model_executor/models/gemma3n_mm.py

+1 -1

@@ -707,7 +707,7 @@ class Gemma3nForConditionalGeneration(

707

708 def forward(

709 self,

710 - input_ids: torch.Tensor,

710 + input_ids: torch.Tensor | None,

711 positions: torch.Tensor,

712 intermediate_tensors: IntermediateTensors | None = None,

713 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/glm4.py

+1 -1

@@ -270,7 +270,7 @@ class Glm4ForCausalLM(nn.Module, SupportsLoRA, SupportsPP):

270

271 def forward(

272 self,

273 - input_ids: torch.Tensor,

273 + input_ids: torch.Tensor | None,

274 positions: torch.Tensor,

275 intermediate_tensors: IntermediateTensors | None = None,

276 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/glm4_1v.py

+1 -1

@@ -1711,7 +1711,7 @@ class Glm4vForConditionalGeneration(

1711

1712 def forward(

1713 self,

1714 - input_ids: torch.Tensor,

1714 + input_ids: torch.Tensor | None,

1715 positions: torch.Tensor,

1716 intermediate_tensors: IntermediateTensors | None = None,

1717 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/glm4_moe.py

+2 -2

@@ -451,7 +451,7 @@ class Glm4MoeModel(nn.Module):

451

452 def forward(

453 self,

454 - input_ids: torch.Tensor,

454 + input_ids: torch.Tensor | None,

455 positions: torch.Tensor,

456 intermediate_tensors: IntermediateTensors | None = None,

457 inputs_embeds: torch.Tensor | None = None,

@@ -687,7 +687,7 @@ class Glm4MoeForCausalLM(nn.Module, SupportsPP, SupportsLoRA, Glm4MixtureOfExper

687

688 def forward(

689 self,

690 - input_ids: torch.Tensor,

690 + input_ids: torch.Tensor | None,

691 positions: torch.Tensor,

692 intermediate_tensors: IntermediateTensors | None = None,

693 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/glm4_moe_lite.py

+2 -2

@@ -264,7 +264,7 @@ class Glm4MoeLiteModel(nn.Module):

264

265 def forward(

266 self,

267 - input_ids: torch.Tensor,

267 + input_ids: torch.Tensor | None,

268 positions: torch.Tensor,

269 intermediate_tensors: IntermediateTensors | None = None,

270 inputs_embeds: torch.Tensor | None = None,

@@ -596,7 +596,7 @@ class Glm4MoeLiteForCausalLM(

596

597 def forward(

598 self,

599 - input_ids: torch.Tensor,

599 + input_ids: torch.Tensor | None,

600 positions: torch.Tensor,

601 intermediate_tensors: IntermediateTensors | None = None,

602 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/glm4_moe_lite_mtp.py

+1 -1

@@ -230,7 +230,7 @@ class Glm4MoeLiteMTP(nn.Module, SupportsPP, Glm4MixtureOfExperts):

230

231 def forward(

232 self,

233 - input_ids: torch.Tensor,

233 + input_ids: torch.Tensor | None,

234 positions: torch.Tensor,

235 hidden_states: torch.Tensor,

236 intermediate_tensors: IntermediateTensors | None = None,

vllm/model_executor/models/glm4_moe_mtp.py

+1 -1

@@ -216,7 +216,7 @@ class Glm4MoeMTP(nn.Module, Glm4MixtureOfExperts):

216

217 def forward(

218 self,

219 - input_ids: torch.Tensor,

219 + input_ids: torch.Tensor | None,

220 positions: torch.Tensor,

221 hidden_states: torch.Tensor,

222 intermediate_tensors: IntermediateTensors | None = None,

vllm/model_executor/models/glm4v.py

+1 -1

@@ -769,7 +769,7 @@ class GLM4VForCausalLM(

769

770 def forward(

771 self,

772 - input_ids: torch.Tensor,

772 + input_ids: torch.Tensor | None,

773 positions: torch.Tensor,

774 intermediate_tensors: IntermediateTensors | None = None,

775 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/glmasr.py

+1 -1

@@ -1075,7 +1075,7 @@ class GlmAsrForConditionalGeneration(

1075

1076 def forward(

1077 self,

1078 - input_ids: torch.Tensor,

1078 + input_ids: torch.Tensor | None,

1079 positions: torch.Tensor,

1080 intermediate_tensors: IntermediateTensors | None = None,

1081 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/gpt2.py

+3 -3

@@ -218,7 +218,7 @@ class GPT2Model(nn.Module):

218

219 def forward(

220 self,

221 - input_ids: torch.Tensor,

221 + input_ids: torch.Tensor | None,

222 position_ids: torch.Tensor,

223 intermediate_tensors: IntermediateTensors | None,

224 inputs_embeds: torch.Tensor | None,

@@ -298,7 +298,7 @@ class GPT2LMHeadModel(nn.Module, SupportsPP):

298

299 def forward(

300 self,

301 - input_ids: torch.Tensor,

301 + input_ids: torch.Tensor | None,

302 positions: torch.Tensor,

303 intermediate_tensors: IntermediateTensors | None = None,

304 inputs_embeds: torch.Tensor | None = None,

@@ -362,7 +362,7 @@ class GPT2ForSequenceClassification(nn.Module, SupportsCrossEncoding):

362

363 def forward(

364 self,

365 - input_ids: torch.Tensor,

365 + input_ids: torch.Tensor | None,

366 positions: torch.Tensor,

367 intermediate_tensors: IntermediateTensors | None = None,

368 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/gpt_bigcode.py

+2 -2

@@ -235,7 +235,7 @@ class GPTBigCodeModel(nn.Module):

235

236 def forward(

237 self,

238 - input_ids: torch.Tensor,

238 + input_ids: torch.Tensor | None,

239 position_ids: torch.Tensor,

240 intermediate_tensors: IntermediateTensors | None,

241 inputs_embeds: torch.Tensor | None = None,

@@ -311,7 +311,7 @@ class GPTBigCodeForCausalLM(nn.Module, SupportsLoRA, SupportsPP):

311

312 def forward(

313 self,

314 - input_ids: torch.Tensor,

314 + input_ids: torch.Tensor | None,

315 positions: torch.Tensor,

316 intermediate_tensors: IntermediateTensors | None = None,

317 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/gpt_j.py

+2 -2

@@ -220,7 +220,7 @@ class GPTJModel(nn.Module):

220

221 def forward(

222 self,

223 - input_ids: torch.Tensor,

223 + input_ids: torch.Tensor | None,

224 position_ids: torch.Tensor,

225 intermediate_tensors: IntermediateTensors | None,

226 inputs_embeds: torch.Tensor | None = None,

@@ -324,7 +324,7 @@ class GPTJForCausalLM(nn.Module, SupportsPP):

324

325 def forward(

326 self,

327 - input_ids: torch.Tensor,

327 + input_ids: torch.Tensor | None,

328 positions: torch.Tensor,

329 intermediate_tensors: IntermediateTensors | None = None,

330 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/gpt_neox.py

+2 -2

@@ -230,7 +230,7 @@ class GPTNeoXModel(nn.Module):

230

231 def forward(

232 self,

233 - input_ids: torch.Tensor,

233 + input_ids: torch.Tensor | None,

234 position_ids: torch.Tensor,

235 intermediate_tensors: IntermediateTensors | None,

236 inputs_embeds: torch.Tensor | None = None,

@@ -318,7 +318,7 @@ class GPTNeoXForCausalLM(nn.Module, SupportsPP):

318

319 def forward(

320 self,

321 - input_ids: torch.Tensor,

321 + input_ids: torch.Tensor | None,

322 positions: torch.Tensor,

323 intermediate_tensors: IntermediateTensors | None = None,

324 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/gpt_oss.py

+2 -2

@@ -275,7 +275,7 @@ class GptOssModel(nn.Module):

275

276 def forward(

277 self,

278 - input_ids: torch.Tensor,

278 + input_ids: torch.Tensor | None,

279 positions: torch.Tensor,

280 intermediate_tensors: IntermediateTensors | None = None,

281 inputs_embeds: torch.Tensor | None = None,

@@ -714,7 +714,7 @@ class GptOssForCausalLM(nn.Module, SupportsPP, SupportsEagle3, SupportsLoRA):

714

715 def forward(

716 self,

717 - input_ids: torch.Tensor,

717 + input_ids: torch.Tensor | None,

718 positions: torch.Tensor,

719 intermediate_tensors: IntermediateTensors | None = None,

720 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/granite.py

+1 -1

@@ -437,7 +437,7 @@ class GraniteForCausalLM(nn.Module, SupportsLoRA, SupportsPP):

437

438 def forward(

439 self,

440 - input_ids: torch.Tensor,

440 + input_ids: torch.Tensor | None,

441 positions: torch.Tensor,

442 intermediate_tensors: IntermediateTensors | None = None,

443 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/granite_speech.py

+1 -1

@@ -806,7 +806,7 @@ class GraniteSpeechForConditionalGeneration(

806

807 def forward(

808 self,

809 - input_ids: torch.Tensor,

809 + input_ids: torch.Tensor | None,

810 positions: torch.Tensor,

811 intermediate_tensors: IntermediateTensors | None = None,

812 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/granitemoe.py

+2 -2

@@ -312,7 +312,7 @@ class GraniteMoeModel(nn.Module):

312

313 def forward(

314 self,

315 - input_ids: torch.Tensor,

315 + input_ids: torch.Tensor | None,

316 positions: torch.Tensor,

317 intermediate_tensors: IntermediateTensors | None,

318 inputs_embeds: torch.Tensor | None = None,

@@ -528,7 +528,7 @@ class GraniteMoeForCausalLM(nn.Module, SupportsLoRA, SupportsPP):

528

529 def forward(

530 self,

531 - input_ids: torch.Tensor,

531 + input_ids: torch.Tensor | None,

532 positions: torch.Tensor,

533 intermediate_tensors: IntermediateTensors | None = None,

534 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/granitemoehybrid.py

+2 -2

@@ -368,7 +368,7 @@ class GraniteMoeHybridModel(nn.Module):

368

369 def forward(

370 self,

371 - input_ids: torch.Tensor,

371 + input_ids: torch.Tensor | None,

372 positions: torch.Tensor,

373 intermediate_tensors: IntermediateTensors | None = None,

374 inputs_embeds: torch.Tensor | None = None,

@@ -685,7 +685,7 @@ class GraniteMoeHybridForCausalLM(

685

686 def forward(

687 self,

688 - input_ids: torch.Tensor,

688 + input_ids: torch.Tensor | None,

689 positions: torch.Tensor,

690 intermediate_tensors: IntermediateTensors | None = None,

691 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/granitemoeshared.py

+2 -2

@@ -183,7 +183,7 @@ class GraniteMoeSharedModel(nn.Module):

183

184 def forward(

185 self,

186 - input_ids: torch.Tensor,

186 + input_ids: torch.Tensor | None,

187 positions: torch.Tensor,

188 intermediate_tensors: IntermediateTensors | None,

189 inputs_embeds: torch.Tensor | None = None,

@@ -295,7 +295,7 @@ class GraniteMoeSharedForCausalLM(nn.Module, SupportsLoRA, SupportsPP):

295

296 def forward(

297 self,

298 - input_ids: torch.Tensor,

298 + input_ids: torch.Tensor | None,

299 positions: torch.Tensor,

300 intermediate_tensors: IntermediateTensors | None = None,

301 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/grok1.py

+2 -2

@@ -491,7 +491,7 @@ class Grok1Model(nn.Module):

491

492 def forward(

493 self,

494 - input_ids: torch.Tensor,

494 + input_ids: torch.Tensor | None,

495 positions: torch.Tensor,

496 intermediate_tensors: IntermediateTensors | None,

497 inputs_embeds: torch.Tensor | None = None,

@@ -705,7 +705,7 @@ class GrokBaseForCausalLM(nn.Module, SupportsLoRA, SupportsPP):

705

706 def forward(

707 self,

708 - input_ids: torch.Tensor,

708 + input_ids: torch.Tensor | None,

709 positions: torch.Tensor,

710 intermediate_tensors: IntermediateTensors | None = None,

711 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/hunyuan_v1.py

+1 -1

@@ -938,7 +938,7 @@ class HunyuanV1ModelBase(nn.Module, SupportsLoRA, SupportsPP):

938

939 def forward(

940 self,

941 - input_ids: torch.Tensor,

941 + input_ids: torch.Tensor | None,

942 positions: torch.Tensor,

943 intermediate_tensors: IntermediateTensors | None = None,

944 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/hunyuan_vision.py

+1 -1

@@ -968,7 +968,7 @@ class HunYuanVLForConditionalGeneration(

968

969 def forward(

970 self,

971 - input_ids: torch.Tensor,

971 + input_ids: torch.Tensor | None,

972 positions: torch.Tensor,

973 intermediate_tensors: IntermediateTensors | None,

974 inputs_embeds: torch.Tensor | None,

vllm/model_executor/models/hyperclovax_vision.py

+1 -1

@@ -747,7 +747,7 @@ class HCXVisionForCausalLM(nn.Module, SupportsMultiModal, SupportsPP):

747

748 def forward(

749 self,

750 - input_ids: torch.Tensor,

750 + input_ids: torch.Tensor | None,

751 positions: torch.Tensor,

752 intermediate_tensors: IntermediateTensors | None = None,

753 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/idefics3.py

+2 -2

@@ -559,7 +559,7 @@ class Idefics3Model(nn.Module):

559

560 def forward(

561 self,

562 - input_ids: torch.Tensor,

562 + input_ids: torch.Tensor | None,

563 positions: torch.Tensor,

564 intermediate_tensors: IntermediateTensors | None = None,

565 inputs_embeds: torch.Tensor | None = None,

@@ -689,7 +689,7 @@ class Idefics3ForConditionalGeneration(nn.Module, SupportsMultiModal, SupportsLo

689

690 def forward(

691 self,

692 - input_ids: torch.Tensor,

692 + input_ids: torch.Tensor | None,

693 positions: torch.Tensor,

694 intermediate_tensors: IntermediateTensors | None = None,

695 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/interfaces.py

+4

@@ -603,6 +603,8 @@ class SupportsPP(Protocol):

603

604 def forward(

605 self,

606 + input_ids: Tensor | None,

607 + positions: Tensor,

608 *,

609 intermediate_tensors: IntermediateTensors | None,

610 ) -> IntermediateTensors | None:

@@ -631,6 +633,8 @@ class _SupportsPPType(Protocol):

633

634 def forward(

635 self,

636 + input_ids: Tensor | None,

637 + positions: Tensor,

638 *,

639 intermediate_tensors: IntermediateTensors | None,

640 ) -> Tensor | IntermediateTensors: ...

vllm/model_executor/models/internlm2.py

+3 -3

@@ -284,7 +284,7 @@ class InternLM2Model(nn.Module):

284

285 def forward(

286 self,

287 - input_ids: torch.Tensor,

287 + input_ids: torch.Tensor | None,

288 positions: torch.Tensor,

289 intermediate_tensors: IntermediateTensors | None = None,

290 inputs_embeds: torch.Tensor | None = None,

@@ -350,7 +350,7 @@ class InternLM2ForCausalLM(nn.Module, SupportsPP, SupportsLoRA):

350

351 def forward(

352 self,

353 - input_ids: torch.Tensor,

353 + input_ids: torch.Tensor | None,

354 positions: torch.Tensor,

355 intermediate_tensors: IntermediateTensors | None,

356 inputs_embeds: torch.Tensor | None = None,

@@ -446,7 +446,7 @@ class InternLM2ForRewardModel(InternLM2ForCausalLM):

446

447 def forward(

448 self,

449 - input_ids: torch.Tensor,

449 + input_ids: torch.Tensor | None,

450 positions: torch.Tensor,

451 intermediate_tensors: IntermediateTensors | None = None,

452 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/internlm2_ve.py

+1 -1

@@ -101,7 +101,7 @@ class InternLM2VEModel(InternLM2Model):

101

102 def forward(

103 self,

104 - input_ids: torch.Tensor,

104 + input_ids: torch.Tensor | None,

105 positions: torch.Tensor,

106 intermediate_tensors: IntermediateTensors | None = None,

107 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/interns1.py

+1 -1

@@ -782,7 +782,7 @@ class InternS1ForConditionalGeneration(

782

783 def forward(

784 self,

785 - input_ids: torch.Tensor,

785 + input_ids: torch.Tensor | None,

786 positions: torch.Tensor,

787 intermediate_tensors: IntermediateTensors | None = None,

788 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/internvl.py

+1 -1

@@ -1371,7 +1371,7 @@ class InternVLChatModel(nn.Module, SupportsMultiModal, SupportsPP, SupportsLoRA)

1371

1372 def forward(

1373 self,

1374 - input_ids: torch.Tensor,

1374 + input_ids: torch.Tensor | None,

1375 positions: torch.Tensor,

1376 intermediate_tensors: IntermediateTensors | None = None,

1377 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/iquest_loopcoder.py

+2 -2

@@ -438,7 +438,7 @@ class IQuestLoopCoderModel(nn.Module):

438

439 def forward(

440 self,

441 - input_ids: torch.Tensor,

441 + input_ids: torch.Tensor | None,

442 positions: torch.Tensor,

443 intermediate_tensors: IntermediateTensors | None = None,

444 inputs_embeds: torch.Tensor | None = None,

@@ -570,7 +570,7 @@ class IQuestLoopCoderForCausalLM(nn.Module):

570

571 def forward(

572 self,

573 - input_ids: torch.Tensor,

573 + input_ids: torch.Tensor | None,

574 positions: torch.Tensor,

575 intermediate_tensors: IntermediateTensors | None = None,

576 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/isaac.py

+1 -1

@@ -1450,7 +1450,7 @@ class IsaacForConditionalGeneration(

1450

1451 def forward(

1452 self,

1453 - input_ids: torch.Tensor,

1453 + input_ids: torch.Tensor | None,

1454 positions: torch.Tensor,

1455 intermediate_tensors: IntermediateTensors | None = None,

1456 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/jais.py

+2 -2

@@ -280,7 +280,7 @@ class JAISModel(nn.Module):

280

281 def forward(

282 self,

283 - input_ids: torch.Tensor,

283 + input_ids: torch.Tensor | None,

284 position_ids: torch.Tensor,

285 intermediate_tensors: IntermediateTensors | None = None,

286 inputs_embeds: torch.Tensor | None = None,

@@ -344,7 +344,7 @@ class JAISLMHeadModel(nn.Module, SupportsPP):

344

345 def forward(

346 self,

347 - input_ids: torch.Tensor,

347 + input_ids: torch.Tensor | None,

348 positions: torch.Tensor,

349 intermediate_tensors: IntermediateTensors | None = None,

350 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/jais2.py

+1 -1

@@ -483,7 +483,7 @@ class Jais2ForCausalLM(nn.Module, SupportsLoRA, SupportsPP):

483

484 def forward(

485 self,

486 - input_ids: torch.Tensor,

486 + input_ids: torch.Tensor | None,

487 positions: torch.Tensor,

488 intermediate_tensors: IntermediateTensors | None = None,

489 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/jamba.py

+2 -2

@@ -348,7 +348,7 @@ class JambaModel(nn.Module):

348

349 def forward(

350 self,

351 - input_ids: torch.Tensor,

351 + input_ids: torch.Tensor | None,

352 positions: torch.Tensor,

353 intermediate_tensors: IntermediateTensors | None = None,

354 inputs_embeds: torch.Tensor | None = None,

@@ -516,7 +516,7 @@ class JambaForCausalLM(

516

517 def forward(

518 self,

519 - input_ids: torch.Tensor,

519 + input_ids: torch.Tensor | None,

520 positions: torch.Tensor,

521 intermediate_tensors: IntermediateTensors | None = None,

522 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/jina_vl.py

+1 -1

@@ -125,7 +125,7 @@ class JinaVLForSequenceClassification(

125

126 def forward(

127 self,

128 - input_ids: torch.Tensor,

128 + input_ids: torch.Tensor | None,

129 positions: torch.Tensor,

130 intermediate_tensors: IntermediateTensors | None = None,

131 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/kanana_v.py

+1 -1

@@ -732,7 +732,7 @@ class KananaVForConditionalGeneration(nn.Module, SupportsMultiModal, SupportsPP)

732

733 def forward(

734 self,

735 - input_ids: torch.Tensor,

735 + input_ids: torch.Tensor | None,

736 positions: torch.Tensor,

737 intermediate_tensors: IntermediateTensors | None = None,

738 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/keye.py

+1 -1

@@ -1438,7 +1438,7 @@ class BaseKeyeModule(nn.Module, SupportsMultiModal):

1438

1439 def forward(

1440 self,

1441 - input_ids: torch.Tensor,

1441 + input_ids: torch.Tensor | None,

1442 positions: torch.Tensor,

1443 intermediate_tensors: IntermediateTensors | None = None,

1444 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/kimi_linear.py

+1 -1

@@ -506,7 +506,7 @@ class KimiLinearForCausalLM(

506

507 def forward(

508 self,

509 - input_ids: torch.Tensor,

509 + input_ids: torch.Tensor | None,

510 positions: torch.Tensor,

511 intermediate_tensors: IntermediateTensors | None = None,

512 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/kimi_vl.py

+1 -1

@@ -389,7 +389,7 @@ class KimiVLForConditionalGeneration(nn.Module, SupportsMultiModal, SupportsPP):

389

390 def forward(

391 self,

392 - input_ids: torch.Tensor,

392 + input_ids: torch.Tensor | None,

393 positions: torch.Tensor,

394 intermediate_tensors: IntermediateTensors | None = None,

395 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/lfm2.py

+2 -2

@@ -342,7 +342,7 @@ class Lfm2Model(nn.Module):

342

343 def forward(

344 self,

345 - input_ids: torch.Tensor,

345 + input_ids: torch.Tensor | None,

346 positions: torch.Tensor,

347 intermediate_tensors: IntermediateTensors | None = None,

348 inputs_embeds: torch.Tensor | None = None,

@@ -503,7 +503,7 @@ class Lfm2ForCausalLM(

503

504 def forward(

505 self,

506 - input_ids: torch.Tensor,

506 + input_ids: torch.Tensor | None,

507 positions: torch.Tensor,

508 intermediate_tensors: IntermediateTensors | None = None,

509 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/lfm2_moe.py

+2 -2

@@ -457,7 +457,7 @@ class Lfm2MoeModel(nn.Module):

457

458 def forward(

459 self,

460 - input_ids: torch.Tensor,

460 + input_ids: torch.Tensor | None,

461 positions: torch.Tensor,

462 intermediate_tensors: IntermediateTensors | None = None,

463 inputs_embeds: torch.Tensor | None = None,

@@ -730,7 +730,7 @@ class Lfm2MoeForCausalLM(

730

731 def forward(

732 self,

733 - input_ids: torch.Tensor,

733 + input_ids: torch.Tensor | None,

734 positions: torch.Tensor,

735 intermediate_tensors: IntermediateTensors | None = None,

736 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/lfm2_vl.py

+1 -1

@@ -769,7 +769,7 @@ class Lfm2VLForConditionalGeneration(

769

770 def forward(

771 self,

772 - input_ids: torch.Tensor,

772 + input_ids: torch.Tensor | None,

773 positions: torch.Tensor,

774 intermediate_tensors: IntermediateTensors | None = None,

775 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/llama.py

+1 -1

@@ -581,7 +581,7 @@ class LlamaForCausalLM(

581

582 def forward(

583 self,

584 - input_ids: torch.Tensor,

584 + input_ids: torch.Tensor | None,

585 positions: torch.Tensor,

586 intermediate_tensors: IntermediateTensors | None = None,

587 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/llava.py

+1 -1

@@ -662,7 +662,7 @@ class LlavaForConditionalGeneration(

662

663 def forward(

664 self,

665 - input_ids: torch.Tensor,

665 + input_ids: torch.Tensor | None,

666 positions: torch.Tensor,

667 intermediate_tensors: IntermediateTensors | None = None,

668 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/llava_next.py

+1 -1

@@ -509,7 +509,7 @@ class LlavaNextForConditionalGeneration(nn.Module, SupportsMultiModal, SupportsP

509

510 def forward(

511 self,

512 - input_ids: torch.Tensor,

512 + input_ids: torch.Tensor | None,

513 positions: torch.Tensor,

514 intermediate_tensors: IntermediateTensors | None = None,

515 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/llava_next_video.py

+1 -1

@@ -426,7 +426,7 @@ class LlavaNextVideoForConditionalGeneration(nn.Module, SupportsMultiModal, Supp

426

427 def forward(

428 self,

429 - input_ids: torch.Tensor,

429 + input_ids: torch.Tensor | None,

430 positions: torch.Tensor,

431 intermediate_tensors: IntermediateTensors | None = None,

432 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/llava_onevision.py

+1 -1

@@ -887,7 +887,7 @@ class LlavaOnevisionForConditionalGeneration(nn.Module, SupportsMultiModal, Supp

887

888 def forward(

889 self,

890 - input_ids: torch.Tensor,

890 + input_ids: torch.Tensor | None,

891 positions: torch.Tensor,

892 intermediate_tensors: IntermediateTensors | None = None,

893 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/longcat_flash.py

+2 -2

@@ -520,7 +520,7 @@ class FlashModel(nn.Module):

520

521 def forward(

522 self,

523 - input_ids: torch.Tensor,

523 + input_ids: torch.Tensor | None,

524 positions: torch.Tensor,

525 intermediate_tensors: IntermediateTensors | None = None,

526 inputs_embeds: torch.Tensor | None = None,

@@ -605,7 +605,7 @@ class LongcatFlashForCausalLM(nn.Module, SupportsLoRA, SupportsPP):

605

606 def forward(

607 self,

608 - input_ids: torch.Tensor,

608 + input_ids: torch.Tensor | None,

609 positions: torch.Tensor,

610 intermediate_tensors: IntermediateTensors | None = None,

611 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/longcat_flash_mtp.py

+1 -1

@@ -150,7 +150,7 @@ class LongCatFlashMTP(nn.Module):

150

151 def forward(

152 self,

153 - input_ids: torch.Tensor,

153 + input_ids: torch.Tensor | None,

154 positions: torch.Tensor,

155 hidden_states: torch.Tensor,

156 intermediate_tensors: IntermediateTensors | None = None,

vllm/model_executor/models/mamba.py

+2 -2

@@ -142,7 +142,7 @@ class MambaModel(nn.Module):

142

143 def forward(

144 self,

145 - input_ids: torch.Tensor,

145 + input_ids: torch.Tensor | None,

146 positions: torch.Tensor,

147 intermediate_tensors: IntermediateTensors | None = None,

148 inputs_embeds: torch.Tensor | None = None,

@@ -225,7 +225,7 @@ class MambaForCausalLM(

225

226 def forward(

227 self,

228 - input_ids: torch.Tensor,

228 + input_ids: torch.Tensor | None,

229 positions: torch.Tensor,

230 intermediate_tensors: IntermediateTensors | None = None,

231 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/mamba2.py

+2 -2

@@ -137,7 +137,7 @@ class Mamba2Model(nn.Module):

137

138 def forward(

139 self,

140 - input_ids: torch.Tensor,

140 + input_ids: torch.Tensor | None,

141 positions: torch.Tensor,

142 intermediate_tensors: IntermediateTensors | None = None,

143 inputs_embeds: torch.Tensor | None = None,

@@ -267,7 +267,7 @@ class Mamba2ForCausalLM(

267

268 def forward(

269 self,

270 - input_ids: torch.Tensor,

270 + input_ids: torch.Tensor | None,

271 positions: torch.Tensor,

272 intermediate_tensors: IntermediateTensors | None = None,

273 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/midashenglm.py

+1 -1

@@ -796,7 +796,7 @@ class MiDashengLMModel(nn.Module, SupportsMultiModal, SupportsPP):

796

797 def forward(

798 self,

799 - input_ids: torch.Tensor,

799 + input_ids: torch.Tensor | None,

800 positions: torch.Tensor,

801 intermediate_tensors: IntermediateTensors | None = None,

802 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/mimo.py

+1 -1

@@ -61,7 +61,7 @@ logger = init_logger(__name__)

61 class MiMoModel(Qwen2Model):

62 def forward(

63 self,

64 - input_ids: torch.Tensor,

64 + input_ids: torch.Tensor | None,

65 positions: torch.Tensor,

66 intermediate_tensors: IntermediateTensors | None = None,

67 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/mimo_mtp.py

+1 -1

@@ -169,7 +169,7 @@ class MiMoMTP(nn.Module):

169

170 def forward(

171 self,

172 - input_ids: torch.Tensor,

172 + input_ids: torch.Tensor | None,

173 positions: torch.Tensor,

174 hidden_states: torch.Tensor,

175 intermediate_tensors: IntermediateTensors | None = None,

vllm/model_executor/models/mimo_v2_flash.py

+2 -2

@@ -478,7 +478,7 @@ class MiMoV2Model(nn.Module):

478

479 def forward(

480 self,

481 - input_ids: torch.Tensor,

481 + input_ids: torch.Tensor | None,

482 positions: torch.Tensor,

483 intermediate_tensors: IntermediateTensors | None = None,

484 inputs_embeds: torch.Tensor | None = None,

@@ -693,7 +693,7 @@ class MiMoV2FlashForCausalLM(nn.Module, SupportsPP, MixtureOfExperts):

693

694 def forward(

695 self,

696 - input_ids: torch.Tensor,

696 + input_ids: torch.Tensor | None,

697 positions: torch.Tensor,

698 intermediate_tensors: IntermediateTensors | None = None,

699 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/minicpm.py

+2 -2

@@ -440,7 +440,7 @@ class MiniCPMModel(nn.Module):

440

441 def forward(

442 self,

443 - input_ids: torch.Tensor,

443 + input_ids: torch.Tensor | None,

444 positions: torch.Tensor,

445 intermediate_tensors: IntermediateTensors | None = None,

446 inputs_embeds: torch.Tensor | None = None,

@@ -620,7 +620,7 @@ class MiniCPMForCausalLM(nn.Module, SupportsLoRA, SupportsPP, SupportsEagle3):

620

621 def forward(

622 self,

623 - input_ids: torch.Tensor,

623 + input_ids: torch.Tensor | None,

624 positions: torch.Tensor,

625 intermediate_tensors: IntermediateTensors | None = None,

626 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/minicpmv.py

+1 -1

@@ -1147,7 +1147,7 @@ class MiniCPMVBaseModel(nn.Module, SupportsMultiModal, SupportsPP):

1147

1148 def forward(

1149 self,

1150 - input_ids: torch.Tensor,

1150 + input_ids: torch.Tensor | None,

1151 positions: torch.Tensor,

1152 intermediate_tensors: IntermediateTensors | None = None,

1153 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/minimax_m2.py

+2 -2

@@ -361,7 +361,7 @@ class MiniMaxM2Model(nn.Module):

361

362 def forward(

363 self,

364 - input_ids: torch.Tensor,

364 + input_ids: torch.Tensor | None,

365 positions: torch.Tensor,

366 intermediate_tensors: IntermediateTensors | None,

367 inputs_embeds: torch.Tensor | None = None,

@@ -520,7 +520,7 @@ class MiniMaxM2ForCausalLM(nn.Module, SupportsLoRA, SupportsPP):

520

521 def forward(

522 self,

523 - input_ids: torch.Tensor,

523 + input_ids: torch.Tensor | None,

524 positions: torch.Tensor,

525 intermediate_tensors: IntermediateTensors | None = None,

526 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/minimax_text_01.py

+1 -1

@@ -712,7 +712,7 @@ class MiniMaxText01ForCausalLM(nn.Module, HasInnerState, IsHybrid):

712

713 def forward(

714 self,

715 - input_ids: torch.Tensor,

715 + input_ids: torch.Tensor | None,

716 positions: torch.Tensor,

717 intermediate_tensors: IntermediateTensors | None = None,

718 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/minimax_vl_01.py

+1 -1

@@ -359,7 +359,7 @@ class MiniMaxVL01ForConditionalGeneration(nn.Module, SupportsMultiModal, Support

359

360 def forward(

361 self,

362 - input_ids: torch.Tensor,

362 + input_ids: torch.Tensor | None,

363 positions: torch.Tensor,

364 intermediate_tensors: IntermediateTensors | None = None,

365 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/mistral3.py

+1 -1

@@ -539,7 +539,7 @@ class Mistral3ForConditionalGeneration(

539

540 def forward(

541 self,

542 - input_ids: torch.Tensor,

542 + input_ids: torch.Tensor | None,

543 positions: torch.Tensor,

544 intermediate_tensors: IntermediateTensors | None = None,

545 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/mixtral.py

+2 -2

@@ -338,7 +338,7 @@ class MixtralModel(nn.Module):

338

339 def forward(

340 self,

341 - input_ids: torch.Tensor,

341 + input_ids: torch.Tensor | None,

342 positions: torch.Tensor,

343 intermediate_tensors: IntermediateTensors | None,

344 inputs_embeds: torch.Tensor | None = None,

@@ -574,7 +574,7 @@ class MixtralForCausalLM(nn.Module, SupportsLoRA, SupportsPP, MixtureOfExperts):

574

575 def forward(

576 self,

577 - input_ids: torch.Tensor,

577 + input_ids: torch.Tensor | None,

578 positions: torch.Tensor,

579 intermediate_tensors: IntermediateTensors | None = None,

580 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/mllama4.py

+1 -1

@@ -901,7 +901,7 @@ class Llama4ForConditionalGeneration(

901

902 def forward(

903 self,

904 - input_ids: torch.Tensor,

904 + input_ids: torch.Tensor | None,

905 positions: torch.Tensor,

906 intermediate_tensors: IntermediateTensors | None = None,

907 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/modernbert.py

+4 -5

@@ -54,12 +54,11 @@ class ModernBertEmbeddings(nn.Module):

54 input_ids: torch.Tensor,

55 inputs_embeds: torch.Tensor | None = None,

56 ) -> torch.Tensor:

57 - if inputs_embeds is not None:

58 - return self.norm(inputs_embeds)

59 - else:

57 + if inputs_embeds is None:

58 inputs_embeds = self.tok_embeddings(input_ids)

61 - embeddings = self.norm(inputs_embeds)

62 - return embeddings

59 +

60 + embeddings = self.norm(inputs_embeds)

61 + return embeddings

62

63

64 class ModernBertAttention(nn.Module):

vllm/model_executor/models/molmo.py

+1 -1

@@ -871,7 +871,7 @@ class MolmoModel(nn.Module, SupportsQuant):

871

872 def forward(

873 self,

874 - input_ids: torch.Tensor,

874 + input_ids: torch.Tensor | None,

875 positions: torch.Tensor,

876 intermediate_tensors: IntermediateTensors | None = None,

877 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/molmo2.py

+1 -1

@@ -1217,7 +1217,7 @@ class Molmo2TextModel(nn.Module, SupportsQuant):

1217

1218 def forward(

1219 self,

1220 - input_ids: torch.Tensor,

1220 + input_ids: torch.Tensor | None,

1221 positions: torch.Tensor,

1222 intermediate_tensors: IntermediateTensors | None = None,

1223 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/mpt.py

+2 -2

@@ -253,7 +253,7 @@ class MPTModel(nn.Module):

253

254 def forward(

255 self,

256 - input_ids: torch.Tensor,

256 + input_ids: torch.Tensor | None,

257 position_ids: torch.Tensor,

258 intermediate_tensors: IntermediateTensors | None,

259 inputs_embeds: torch.Tensor | None = None,

@@ -313,7 +313,7 @@ class MPTForCausalLM(nn.Module, SupportsPP):

313

314 def forward(

315 self,

316 - input_ids: torch.Tensor,

316 + input_ids: torch.Tensor | None,

317 positions: torch.Tensor,

318 intermediate_tensors: IntermediateTensors | None = None,

319 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/nano_nemotron_vl.py

+1 -1

@@ -1917,7 +1917,7 @@ class NemotronH_Nano_VL_V2(

1917

1918 def forward(

1919 self,

1920 - input_ids: torch.Tensor,

1920 + input_ids: torch.Tensor | None,

1921 positions: torch.Tensor,

1922 intermediate_tensors: IntermediateTensors | None = None,

1923 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/nemotron.py

+1 -1

@@ -477,7 +477,7 @@ class NemotronForCausalLM(nn.Module, SupportsLoRA, SupportsPP):

477

478 def forward(

479 self,

480 - input_ids: torch.Tensor,

480 + input_ids: torch.Tensor | None,

481 positions: torch.Tensor,

482 intermediate_tensors: IntermediateTensors | None = None,

483 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/nemotron_h.py

+2 -2

@@ -601,7 +601,7 @@ class NemotronHModel(nn.Module):

601

602 def forward(

603 self,

604 - input_ids: torch.Tensor,

604 + input_ids: torch.Tensor | None,

605 positions: torch.Tensor,

606 intermediate_tensors: IntermediateTensors | None = None,

607 inputs_embeds: torch.Tensor | None = None,

@@ -887,7 +887,7 @@ class NemotronHForCausalLM(

887

888 def forward(

889 self,

890 - input_ids: torch.Tensor,

890 + input_ids: torch.Tensor | None,

891 positions: torch.Tensor,

892 intermediate_tensors: IntermediateTensors | None = None,

893 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/nemotron_nas.py

+1 -1

@@ -449,7 +449,7 @@ class DeciLMForCausalLM(nn.Module, SupportsLoRA, SupportsPP, HasNoOps):

449

450 def forward(

451 self,

452 - input_ids: torch.Tensor,

452 + input_ids: torch.Tensor | None,

453 positions: torch.Tensor,

454 intermediate_tensors: IntermediateTensors | None = None,

455 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/nemotron_parse.py

+2 -2

@@ -290,7 +290,7 @@ class MBartDecoderNoPos(nn.Module):

290

291 def forward(

292 self,

293 - decoder_input_ids: torch.Tensor,

293 + decoder_input_ids: torch.Tensor | None,

294 *,

295 encoder_hidden_states: torch.Tensor | None,

296 inputs_embeds: torch.Tensor | None = None,

@@ -894,7 +894,7 @@ class NemotronParseForConditionalGeneration(nn.Module, SupportsMultiModal):

894

895 def forward(

896 self,

897 - input_ids: torch.Tensor,

897 + input_ids: torch.Tensor | None,

898 positions: torch.Tensor,

899 encoder_outputs: list[torch.Tensor] | None = None,

900 **kwargs,

vllm/model_executor/models/nemotron_vl.py

+1 -1

@@ -597,7 +597,7 @@ class LlamaNemotronVLChatModel(nn.Module, SupportsMultiModal, SupportsPP, Suppor

597

598 def forward(

599 self,

600 - input_ids: torch.Tensor,

600 + input_ids: torch.Tensor | None,

601 positions: torch.Tensor,

602 intermediate_tensors: IntermediateTensors | None = None,

603 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/olmo.py

+2 -2

@@ -271,7 +271,7 @@ class OlmoModel(nn.Module):

271

272 def forward(

273 self,

274 - input_ids: torch.Tensor,

274 + input_ids: torch.Tensor | None,

275 positions: torch.Tensor,

276 intermediate_tensors: IntermediateTensors | None,

277 inputs_embeds: torch.Tensor | None = None,

@@ -382,7 +382,7 @@ class OlmoForCausalLM(nn.Module, SupportsPP, SupportsLoRA):

382

383 def forward(

384 self,

385 - input_ids: torch.Tensor,

385 + input_ids: torch.Tensor | None,

386 positions: torch.Tensor,

387 intermediate_tensors: IntermediateTensors | None = None,

388 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/olmo2.py

+2 -2

@@ -309,7 +309,7 @@ class Olmo2Model(nn.Module):

309

310 def forward(

311 self,

312 - input_ids: torch.Tensor,

312 + input_ids: torch.Tensor | None,

313 positions: torch.Tensor,

314 intermediate_tensors: IntermediateTensors | None,

315 inputs_embeds: torch.Tensor | None = None,

@@ -424,7 +424,7 @@ class Olmo2ForCausalLM(nn.Module, SupportsPP, SupportsLoRA):

424

425 def forward(

426 self,

427 - input_ids: torch.Tensor,

427 + input_ids: torch.Tensor | None,

428 positions: torch.Tensor,

429 intermediate_tensors: IntermediateTensors | None = None,

430 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/olmoe.py

+2 -2

@@ -300,7 +300,7 @@ class OlmoeModel(nn.Module):

300

301 def forward(

302 self,

303 - input_ids: torch.Tensor,

303 + input_ids: torch.Tensor | None,

304 positions: torch.Tensor,

305 intermediate_tensors: IntermediateTensors | None,

306 inputs_embeds: torch.Tensor | None = None,

@@ -476,7 +476,7 @@ class OlmoeForCausalLM(nn.Module, SupportsPP, SupportsLoRA):

476

477 def forward(

478 self,

479 - input_ids: torch.Tensor,

479 + input_ids: torch.Tensor | None,

480 positions: torch.Tensor,

481 intermediate_tensors: IntermediateTensors | None = None,

482 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/openpangu.py

+2 -2

@@ -1056,7 +1056,7 @@ class OpenPanguModel(nn.Module):

1056

1057 def forward(

1058 self,

1059 - input_ids: torch.Tensor,

1059 + input_ids: torch.Tensor | None,

1060 positions: torch.Tensor,

1061 intermediate_tensors: IntermediateTensors | None,

1062 inputs_embeds: torch.Tensor | None = None,

@@ -1286,7 +1286,7 @@ class OpenPanguModelBase(nn.Module, SupportsPP, SupportsLoRA):

1286

1287 def forward(

1288 self,

1289 - input_ids: torch.Tensor,

1289 + input_ids: torch.Tensor | None,

1290 positions: torch.Tensor,

1291 intermediate_tensors: IntermediateTensors | None = None,

1292 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/openpangu_mtp.py

+1 -1

@@ -104,7 +104,7 @@ class OpenPanguMTP(nn.Module):

104

105 def forward(

106 self,

107 - input_ids: torch.Tensor,

107 + input_ids: torch.Tensor | None,

108 positions: torch.Tensor,

109 hidden_states: torch.Tensor,

110 intermediate_tensors: IntermediateTensors | None = None,

vllm/model_executor/models/opt.py

+3 -3

@@ -267,7 +267,7 @@ class OPTDecoder(nn.Module):

267

268 def forward(

269 self,

270 - input_ids: torch.Tensor,

270 + input_ids: torch.Tensor | None,

271 positions: torch.Tensor,

272 intermediate_tensors: IntermediateTensors | None,

273 inputs_embeds: torch.Tensor | None = None,

@@ -316,7 +316,7 @@ class OPTModel(nn.Module):

316

317 def forward(

318 self,

319 - input_ids: torch.Tensor,

319 + input_ids: torch.Tensor | None,

320 positions: torch.Tensor,

321 intermediate_tensors: IntermediateTensors | None,

322 inputs_embeds: torch.Tensor | None = None,

@@ -399,7 +399,7 @@ class OPTForCausalLM(nn.Module, SupportsPP, SupportsLoRA):

399

400 def forward(

401 self,

402 - input_ids: torch.Tensor,

402 + input_ids: torch.Tensor | None,

403 positions: torch.Tensor,

404 intermediate_tensors: IntermediateTensors | None = None,

405 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/orion.py

+2 -2

@@ -253,7 +253,7 @@ class OrionModel(nn.Module):

253

254 def forward(

255 self,

256 - input_ids: torch.Tensor,

256 + input_ids: torch.Tensor | None,

257 positions: torch.Tensor,

258 intermediate_tensors: IntermediateTensors | None,

259 inputs_embeds: torch.Tensor | None = None,

@@ -343,7 +343,7 @@ class OrionForCausalLM(nn.Module, SupportsPP):

343

344 def forward(

345 self,

346 - input_ids: torch.Tensor,

346 + input_ids: torch.Tensor | None,

347 positions: torch.Tensor,

348 intermediate_tensors: IntermediateTensors | None = None,

349 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/ouro.py

+2 -2

@@ -357,7 +357,7 @@ class OuroModel(nn.Module):

357

358 def forward(

359 self,

360 - input_ids: torch.Tensor,

360 + input_ids: torch.Tensor | None,

361 positions: torch.Tensor,

362 intermediate_tensors: IntermediateTensors | None = None,

363 inputs_embeds: torch.Tensor | None = None,

@@ -482,7 +482,7 @@ class OuroForCausalLM(nn.Module, SupportsLoRA):

482

483 def forward(

484 self,

485 - input_ids: torch.Tensor,

485 + input_ids: torch.Tensor | None,

486 positions: torch.Tensor,

487 intermediate_tensors: IntermediateTensors | None = None,

488 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/ovis.py

+1 -1

@@ -525,7 +525,7 @@ class Ovis(nn.Module, SupportsMultiModal, SupportsPP):

525

526 def forward(

527 self,

528 - input_ids: torch.Tensor,

528 + input_ids: torch.Tensor | None,

529 positions: torch.Tensor,

530 intermediate_tensors: IntermediateTensors | None = None,

531 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/ovis2_5.py

+1 -1

@@ -632,7 +632,7 @@ class Ovis2_5(nn.Module, SupportsMultiModal, SupportsPP):

632

633 def forward(

634 self,

635 - input_ids: torch.Tensor,

635 + input_ids: torch.Tensor | None,

636 positions: torch.Tensor,

637 intermediate_tensors: IntermediateTensors | None = None,

638 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/paddleocr_vl.py

+1 -1

@@ -1159,7 +1159,7 @@ class PaddleOCRVLForConditionalGeneration(nn.Module, SupportsMultiModal, Support

1159

1160 def forward(

1161 self,

1162 - input_ids: torch.Tensor,

1162 + input_ids: torch.Tensor | None,

1163 positions: torch.Tensor,

1164 intermediate_tensors: IntermediateTensors | None = None,

1165 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/paligemma.py

+1 -1

@@ -389,7 +389,7 @@ class PaliGemmaForConditionalGeneration(

389

390 def forward(

391 self,

392 - input_ids: torch.Tensor,

392 + input_ids: torch.Tensor | None,

393 positions: torch.Tensor,

394 intermediate_tensors: IntermediateTensors | None = None,

395 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/persimmon.py

+2 -2

@@ -271,7 +271,7 @@ class PersimmonModel(nn.Module):

271

272 def forward(

273 self,

274 - input_ids: torch.Tensor,

274 + input_ids: torch.Tensor | None,

275 positions: torch.Tensor,

276 intermediate_tensors: IntermediateTensors | None,

277 inputs_embeds: torch.Tensor | None = None,

@@ -348,7 +348,7 @@ class PersimmonForCausalLM(nn.Module, SupportsPP):

348

349 def forward(

350 self,

351 - input_ids: torch.Tensor,

351 + input_ids: torch.Tensor | None,

352 positions: torch.Tensor,

353 intermediate_tensors: IntermediateTensors | None = None,

354 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/phi.py

+2 -2

@@ -234,7 +234,7 @@ class PhiModel(nn.Module):

234

235 def forward(

236 self,

237 - input_ids: torch.Tensor,

237 + input_ids: torch.Tensor | None,

238 positions: torch.Tensor,

239 intermediate_tensors: IntermediateTensors | None,

240 inputs_embeds: torch.Tensor | None = None,

@@ -340,7 +340,7 @@ class PhiForCausalLM(nn.Module, SupportsLoRA, SupportsPP):

340

341 def forward(

342 self,

343 - input_ids: torch.Tensor,

343 + input_ids: torch.Tensor | None,

344 positions: torch.Tensor,

345 intermediate_tensors: IntermediateTensors | None = None,

346 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/phi3v.py

+1 -1

@@ -686,7 +686,7 @@ class Phi3VForCausalLM(nn.Module, SupportsMultiModal, SupportsPP, SupportsQuant)

686

687 def forward(

688 self,

689 - input_ids: torch.Tensor,

689 + input_ids: torch.Tensor | None,

690 positions: torch.Tensor,

691 intermediate_tensors: IntermediateTensors | None = None,

692 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/phi4mm.py

+1 -1

@@ -1211,7 +1211,7 @@ class Phi4MMForCausalLM(nn.Module, SupportsLoRA, SupportsMultiModal):

1211

1212 def forward(

1213 self,

1214 - input_ids: torch.Tensor,

1214 + input_ids: torch.Tensor | None,

1215 positions: torch.Tensor,

1216 intermediate_tensors: IntermediateTensors | None = None,

1217 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/phimoe.py

+2 -2

@@ -483,7 +483,7 @@ class PhiMoEModel(nn.Module):

483

484 def forward(

485 self,

486 - input_ids: torch.Tensor,

486 + input_ids: torch.Tensor | None,

487 positions: torch.Tensor,

488 intermediate_tensors: IntermediateTensors | None,

489 inputs_embeds: torch.Tensor | None = None,

@@ -649,7 +649,7 @@ class PhiMoEForCausalLM(nn.Module, SupportsLoRA, SupportsPP):

649

650 def forward(

651 self,

652 - input_ids: torch.Tensor,

652 + input_ids: torch.Tensor | None,

653 positions: torch.Tensor,

654 intermediate_tensors: IntermediateTensors | None = None,

655 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/pixtral.py

+1 -1

@@ -479,7 +479,7 @@ class PixtralForConditionalGeneration(

479

480 def forward(

481 self,

482 - input_ids: torch.Tensor,

482 + input_ids: torch.Tensor | None,

483 positions: torch.Tensor,

484 intermediate_tensors: IntermediateTensors | None = None,

485 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/plamo2.py

+2 -2

@@ -775,7 +775,7 @@ class Plamo2Model(torch.nn.Module):

775

776 def forward(

777 self,

778 - input_ids: torch.Tensor,

778 + input_ids: torch.Tensor | None,

779 positions: torch.Tensor,

780 intermediate_tensors: IntermediateTensors | None = None,

781 inputs_embeds: torch.Tensor | None = None,

@@ -852,7 +852,7 @@ class Plamo2ForCausalLM(

852

853 def forward(

854 self,

855 - input_ids: torch.Tensor,

855 + input_ids: torch.Tensor | None,

856 positions: torch.Tensor,

857 intermediate_tensors: IntermediateTensors | None = None,

858 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/plamo3.py

+2 -2

@@ -342,7 +342,7 @@ class Plamo3Model(nn.Module):

342

343 def forward(

344 self,

345 - input_ids: torch.Tensor,

345 + input_ids: torch.Tensor | None,

346 positions: torch.Tensor,

347 intermediate_tensors: IntermediateTensors | None = None,

348 inputs_embeds: torch.Tensor | None = None,

@@ -412,7 +412,7 @@ class Plamo3ForCausalLM(nn.Module, SupportsLoRA, SupportsPP):

412

413 def forward(

414 self,

415 - input_ids: torch.Tensor,

415 + input_ids: torch.Tensor | None,

416 positions: torch.Tensor,

417 intermediate_tensors: IntermediateTensors | None = None,

418 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/qwen.py

+2 -2

@@ -232,7 +232,7 @@ class QWenModel(nn.Module):

232

233 def forward(

234 self,

235 - input_ids: torch.Tensor,

235 + input_ids: torch.Tensor | None,

236 positions: torch.Tensor,

237 intermediate_tensors: IntermediateTensors | None,

238 inputs_embeds: torch.Tensor | None = None,

@@ -366,7 +366,7 @@ class QWenLMHeadModel(QWenBaseModel, SupportsPP, SupportsLoRA):

366

367 def forward(

368 self,

369 - input_ids: torch.Tensor,

369 + input_ids: torch.Tensor | None,

370 positions: torch.Tensor,

371 intermediate_tensors: IntermediateTensors | None = None,

372 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/qwen2.py

+2 -2

@@ -417,7 +417,7 @@ class Qwen2Model(nn.Module):

417

418 def forward(

419 self,

420 - input_ids: torch.Tensor,

420 + input_ids: torch.Tensor | None,

421 positions: torch.Tensor,

422 intermediate_tensors: IntermediateTensors | None = None,

423 inputs_embeds: torch.Tensor | None = None,

@@ -575,7 +575,7 @@ class Qwen2ForCausalLM(nn.Module, SupportsLoRA, SupportsPP, SupportsEagle3):

575

576 def forward(

577 self,

578 - input_ids: torch.Tensor,

578 + input_ids: torch.Tensor | None,

579 positions: torch.Tensor,

580 intermediate_tensors: IntermediateTensors | None = None,

581 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/qwen2_5_omni_thinker.py

+1 -1

@@ -1298,7 +1298,7 @@ class Qwen2_5OmniThinkerForConditionalGeneration(

1298

1299 def forward(

1300 self,

1301 - input_ids: torch.Tensor,

1301 + input_ids: torch.Tensor | None,

1302 positions: torch.Tensor,

1303 intermediate_tensors: IntermediateTensors | None = None,

1304 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/qwen2_5_vl.py

+1 -1

@@ -1448,7 +1448,7 @@ class Qwen2_5_VLForConditionalGeneration(

1448

1449 def forward(

1450 self,

1451 - input_ids: torch.Tensor,

1451 + input_ids: torch.Tensor | None,

1452 positions: torch.Tensor,

1453 intermediate_tensors: IntermediateTensors | None = None,

1454 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/qwen2_audio.py

+1 -1

@@ -451,7 +451,7 @@ class Qwen2AudioForConditionalGeneration(nn.Module, SupportsMultiModal, Supports

451

452 def forward(

453 self,

454 - input_ids: torch.Tensor,

454 + input_ids: torch.Tensor | None,

455 positions: torch.Tensor,

456 intermediate_tensors: IntermediateTensors | None = None,

457 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/qwen2_moe.py

+2 -2

@@ -395,7 +395,7 @@ class Qwen2MoeModel(nn.Module):

395

396 def forward(

397 self,

398 - input_ids: torch.Tensor,

398 + input_ids: torch.Tensor | None,

399 positions: torch.Tensor,

400 intermediate_tensors: IntermediateTensors | None = None,

401 inputs_embeds: torch.Tensor | None = None,

@@ -579,7 +579,7 @@ class Qwen2MoeForCausalLM(nn.Module, SupportsPP, SupportsLoRA):

579

580 def forward(

581 self,

582 - input_ids: torch.Tensor,

582 + input_ids: torch.Tensor | None,

583 positions: torch.Tensor,

584 intermediate_tensors: IntermediateTensors | None = None,

585 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/qwen2_rm.py

+1 -1

@@ -79,7 +79,7 @@ class Qwen2RewardBaseModel(nn.Module, SupportsLoRA, SupportsPP):

79

80 def forward(

81 self,

82 - input_ids: torch.Tensor,

82 + input_ids: torch.Tensor | None,

83 positions: torch.Tensor,

84 intermediate_tensors: IntermediateTensors | None = None,

85 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/qwen2_vl.py

+1 -1

@@ -1368,7 +1368,7 @@ class Qwen2VLForConditionalGeneration(

1368

1369 def forward(

1370 self,

1371 - input_ids: torch.Tensor,

1371 + input_ids: torch.Tensor | None,

1372 positions: torch.Tensor,

1373 intermediate_tensors: IntermediateTensors | None = None,

1374 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/qwen3.py

+1 -1

@@ -306,7 +306,7 @@ class Qwen3ForCausalLM(nn.Module, SupportsLoRA, SupportsPP, SupportsEagle3):

306

307 def forward(

308 self,

309 - input_ids: torch.Tensor,

309 + input_ids: torch.Tensor | None,

310 positions: torch.Tensor,

311 intermediate_tensors: IntermediateTensors | None = None,

312 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/qwen3_moe.py

+2 -2

@@ -464,7 +464,7 @@ class Qwen3MoeModel(nn.Module):

464

465 def forward(

466 self,

467 - input_ids: torch.Tensor,

467 + input_ids: torch.Tensor | None,

468 positions: torch.Tensor,

469 intermediate_tensors: IntermediateTensors | None = None,

470 inputs_embeds: torch.Tensor | None = None,

@@ -765,7 +765,7 @@ class Qwen3MoeForCausalLM(

765

766 def forward(

767 self,

768 - input_ids: torch.Tensor,

768 + input_ids: torch.Tensor | None,

769 positions: torch.Tensor,

770 intermediate_tensors: IntermediateTensors | None = None,

771 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/qwen3_next.py

+2 -2

@@ -1004,7 +1004,7 @@ class Qwen3NextModel(nn.Module):

1004

1005 def forward(

1006 self,

1007 - input_ids: torch.Tensor,

1007 + input_ids: torch.Tensor | None,

1008 positions: torch.Tensor,

1009 intermediate_tensors: IntermediateTensors | None = None,

1010 inputs_embeds: torch.Tensor | None = None,

@@ -1239,7 +1239,7 @@ class Qwen3NextForCausalLM(

1239

1240 def forward(

1241 self,

1242 - input_ids: torch.Tensor,

1242 + input_ids: torch.Tensor | None,

1243 positions: torch.Tensor,

1244 intermediate_tensors: IntermediateTensors | None = None,

1245 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/qwen3_next_mtp.py

+1 -1

@@ -261,7 +261,7 @@ class Qwen3NextMTP(nn.Module, QwenNextMixtureOfExperts):

261

262 def forward(

263 self,

264 - input_ids: torch.Tensor,

264 + input_ids: torch.Tensor | None,

265 positions: torch.Tensor,

266 hidden_states: torch.Tensor,

267 intermediate_tensors: IntermediateTensors | None = None,

vllm/model_executor/models/qwen3_omni_moe_thinker.py

+2 -2

@@ -998,7 +998,7 @@ class Qwen3MoeLLMModel(Qwen3MoeModel):

998

999 def forward(

1000 self,

1001 - input_ids: torch.Tensor,

1001 + input_ids: torch.Tensor | None,

1002 positions: torch.Tensor,

1003 intermediate_tensors: IntermediateTensors | None = None,

1004 inputs_embeds: torch.Tensor | None = None,

@@ -1819,7 +1819,7 @@ class Qwen3OmniMoeThinkerForConditionalGeneration(

1819

1820 def forward(

1821 self,

1822 - input_ids: torch.Tensor,

1822 + input_ids: torch.Tensor | None,

1823 positions: torch.Tensor,

1824 intermediate_tensors: IntermediateTensors | None = None,

1825 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/qwen3_vl.py

+2 -2

@@ -1119,7 +1119,7 @@ class Qwen3LLMModel(Qwen3Model):

1119

1120 def forward(

1121 self,

1122 - input_ids: torch.Tensor,

1122 + input_ids: torch.Tensor | None,

1123 positions: torch.Tensor,

1124 intermediate_tensors: IntermediateTensors | None = None,

1125 inputs_embeds: torch.Tensor | None = None,

@@ -2001,7 +2001,7 @@ class Qwen3VLForConditionalGeneration(

2001

2002 def forward(

2003 self,

2004 - input_ids: torch.Tensor,

2004 + input_ids: torch.Tensor | None,

2005 positions: torch.Tensor,

2006 intermediate_tensors: IntermediateTensors | None = None,

2007 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/qwen3_vl_moe.py

+1 -1

@@ -94,7 +94,7 @@ class Qwen3MoeLLMModel(Qwen3MoeModel):

94

95 def forward(

96 self,

97 - input_ids: torch.Tensor,

97 + input_ids: torch.Tensor | None,

98 positions: torch.Tensor,

99 intermediate_tensors: IntermediateTensors | None = None,

100 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/qwen_vl.py

+1 -1

@@ -810,7 +810,7 @@ class QwenVLForConditionalGeneration(

810

811 def forward(

812 self,

813 - input_ids: torch.Tensor,

813 + input_ids: torch.Tensor | None,

814 positions: torch.Tensor,

815 intermediate_tensors: IntermediateTensors | None = None,

816 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/seed_oss.py

+2 -2

@@ -334,7 +334,7 @@ class SeedOssModel(nn.Module):

334

335 def forward(

336 self,

337 - input_ids: torch.Tensor,

337 + input_ids: torch.Tensor | None,

338 positions: torch.Tensor,

339 intermediate_tensors: IntermediateTensors | None = None,

340 inputs_embeds: torch.Tensor | None = None,

@@ -467,7 +467,7 @@ class SeedOssForCausalLM(nn.Module, SupportsLoRA, SupportsPP):

467

468 def forward(

469 self,

470 - input_ids: torch.Tensor,

470 + input_ids: torch.Tensor | None,

471 positions: torch.Tensor,

472 intermediate_tensors: IntermediateTensors | None = None,

473 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/skyworkr1v.py

+1 -1

@@ -898,7 +898,7 @@ class SkyworkR1VChatModel(nn.Module, SupportsMultiModal, SupportsPP):

898

899 def forward(

900 self,

901 - input_ids: torch.Tensor,

901 + input_ids: torch.Tensor | None,

902 positions: torch.Tensor,

903 intermediate_tensors: IntermediateTensors | None = None,

904 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/solar.py

+1 -1

@@ -465,7 +465,7 @@ class SolarForCausalLM(nn.Module, SupportsLoRA, SupportsPP):

465

466 def forward(

467 self,

468 - input_ids: torch.Tensor,

468 + input_ids: torch.Tensor | None,

469 positions: torch.Tensor,

470 intermediate_tensors: IntermediateTensors | None = None,

471 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/stablelm.py

+2 -2

@@ -246,7 +246,7 @@ class StableLMEpochModel(nn.Module):

246

247 def forward(

248 self,

249 - input_ids: torch.Tensor,

249 + input_ids: torch.Tensor | None,

250 positions: torch.Tensor,

251 intermediate_tensors: IntermediateTensors | None,

252 inputs_embeds: torch.Tensor | None = None,

@@ -332,7 +332,7 @@ class StablelmForCausalLM(nn.Module, SupportsPP):

332

333 def forward(

334 self,

335 - input_ids: torch.Tensor,

335 + input_ids: torch.Tensor | None,

336 positions: torch.Tensor,

337 intermediate_tensors: IntermediateTensors | None = None,

338 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/starcoder2.py

+2 -2

@@ -252,7 +252,7 @@ class Starcoder2Model(nn.Module):

252

253 def forward(

254 self,

255 - input_ids: torch.Tensor,

255 + input_ids: torch.Tensor | None,

256 positions: torch.Tensor,

257 intermediate_tensors: IntermediateTensors | None,

258 inputs_embeds: torch.Tensor | None = None,

@@ -336,7 +336,7 @@ class Starcoder2ForCausalLM(nn.Module, SupportsPP):

336

337 def forward(

338 self,

339 - input_ids: torch.Tensor,

339 + input_ids: torch.Tensor | None,

340 positions: torch.Tensor,

341 intermediate_tensors: IntermediateTensors | None = None,

342 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/step3_text.py

+2 -2

@@ -354,7 +354,7 @@ class Step3TextModel(nn.Module):

354

355 def forward(

356 self,

357 - input_ids: torch.Tensor,

357 + input_ids: torch.Tensor | None,

358 positions: torch.Tensor,

359 intermediate_tensors: IntermediateTensors | None = None,

360 inputs_embeds: torch.Tensor | None = None,

@@ -419,7 +419,7 @@ class Step3TextForCausalLM(nn.Module, SupportsPP):

419

420 def forward(

421 self,

422 - input_ids: torch.Tensor,

422 + input_ids: torch.Tensor | None,

423 positions: torch.Tensor,

424 intermediate_tensors: IntermediateTensors | None = None,

425 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/step3_vl.py

+1 -1

@@ -1101,7 +1101,7 @@ class Step3VLForConditionalGeneration(nn.Module, SupportsMultiModal, SupportsPP)

1101

1102 def forward(

1103 self,

1104 - input_ids: torch.Tensor,

1104 + input_ids: torch.Tensor | None,

1105 positions: torch.Tensor,

1106 intermediate_tensors: IntermediateTensors | None = None,

1107 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/tarsier.py

+1 -1

@@ -585,7 +585,7 @@ class TarsierForConditionalGeneration(nn.Module, SupportsMultiModal, SupportsPP)

585

586 def forward(

587 self,

588 - input_ids: torch.Tensor,

588 + input_ids: torch.Tensor | None,

589 positions: torch.Tensor,

590 intermediate_tensors: IntermediateTensors | None = None,

591 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/ultravox.py

+1 -1

@@ -714,7 +714,7 @@ class UltravoxModel(nn.Module, SupportsMultiModal, SupportsPP, SupportsLoRA):

714

715 def forward(

716 self,

717 - input_ids: torch.Tensor,

717 + input_ids: torch.Tensor | None,

718 positions: torch.Tensor,

719 intermediate_tensors: torch.Tensor | None = None,

720 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/voxtral.py

+1 -1

@@ -397,7 +397,7 @@ class VoxtralForConditionalGeneration(

397

398 def forward(

399 self,

400 - input_ids: torch.Tensor,

400 + input_ids: torch.Tensor | None,

401 positions: torch.Tensor,

402 intermediate_tensors: IntermediateTensors | None = None,

403 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/voxtral_streaming.py

+1 -1

@@ -173,7 +173,7 @@ class VoxtralStreamingGeneration(VoxtralForConditionalGeneration):

173

174 def forward(

175 self,

176 - input_ids: torch.Tensor,

176 + input_ids: torch.Tensor | None,

177 positions: torch.Tensor,

178 intermediate_tensors: IntermediateTensors | None = None,

179 inputs_embeds: torch.Tensor | None = None,

vllm/model_executor/models/zamba2.py

+2 -2

@@ -771,7 +771,7 @@ class Zamba2Model(nn.Module):

771

772 def forward(

773 self,

774 - input_ids: torch.Tensor,

774 + input_ids: torch.Tensor | None,

775 positions: torch.Tensor,

776 inputs_embeds: torch.Tensor | None = None,

777 ) -> torch.Tensor | IntermediateTensors:

@@ -947,7 +947,7 @@ class Zamba2ForCausalLM(nn.Module, HasInnerState, IsHybrid, SupportsMambaPrefixC

947

948 def forward(

949 self,

950 - input_ids: torch.Tensor,

950 + input_ids: torch.Tensor | None,

951 positions: torch.Tensor,

952 inputs_embeds: torch.Tensor | None = None,

953 **kwargs: Any,