A high-throughput and memory-efficient inference and serving engine for LLMs
[Model] Remove InternLMForCausalLM registry alias (#45128)
Signed-off-by: Xianbao QIAN <xianbao.qian@gmail.com> Co-authored-by: Claude <noreply@anthropic.com>
Tiezhen WANG committed
Jun 12, 2026 at 04:02 UTC
c9340e6f350a009cf835878abad2a0e379b9e6a4
14 files changed
+1
-17
docs/models/supported_models.md
-1
@@ -423,7 +423,6 @@ th {
423
| `HunYuanMoEV1ForCausalLM` | Hunyuan-A13B | `tencent/Hunyuan-A13B-Instruct`, `tencent/Hunyuan-A13B-Pretrain`, `tencent/Hunyuan-A13B-Instruct-FP8`, etc. | ✅︎ | ✅︎ |
424
| `HYV3ForCausalLM` | HY3 | `tencent/Hy3-preview-Base`, `tencent/Hy3-preview` | ✅︎ | ✅︎ |
425
| `HyperCLOVAXForCausalLM` | HyperCLOVAX-SEED-Think-14B | `naver-hyperclovax/HyperCLOVAX-SEED-Think-14B` | ✅︎ | ✅︎ |
426
-| `InternLMForCausalLM` | InternLM | `internlm/internlm-7b`, `internlm/internlm-chat-7b`, etc. | ✅︎ | ✅︎ |
426
| `InternLM2ForCausalLM` | InternLM2 | `internlm/internlm2-7b`, `internlm/internlm2-chat-7b`, etc. | ✅︎ | ✅︎ |
427
| `InternLM3ForCausalLM` | InternLM3 | `internlm/internlm3-8b-instruct`, etc. | ✅︎ | ✅︎ |
428
| `IQuestCoderForCausalLM` | IQuestCoderV1 | `IQuestLab/IQuest-Coder-V1-40B-Instruct`, etc. | | |
tests/distributed/test_pipeline_parallel.py
-2
@@ -124,8 +124,6 @@ TEXT_GENERATION_MODELS = {
124
"EleutherAI/pythia-1.4b": PPTestSettings.fast(),
125
"ibm/PowerLM-3b": PPTestSettings.fast(),
126
"ibm/PowerMoE-3b": PPTestSettings.fast(),
127
- # Uses Llama
128
- # "internlm/internlm-chat-7b": PPTestSettings.fast(),
127
"internlm/internlm2-chat-7b": PPTestSettings.fast(),
128
"ai21labs/Jamba-tiny-dev": PPTestSettings.fast(),
129
"pfnet/plamo-2-1b": PPTestSettings.fast(),
tests/models/registry.py
-3
@@ -338,9 +338,6 @@ _TEXT_GENERATION_EXAMPLE_MODELS = {
338
"naver-hyperclovax/HyperCLOVAX-SEED-Think-14B",
339
min_transformers_version="5.9.0",
340
),
341
- "InternLMForCausalLM": _HfExamplesInfo(
342
- "internlm/internlm-chat-7b", trust_remote_code=True
343
- ),
341
"InternLM2ForCausalLM": _HfExamplesInfo(
342
"internlm/internlm2-chat-7b", trust_remote_code=True
343
),
vllm/model_executor/models/apertus.py
-1
@@ -252,7 +252,6 @@ class ApertusDecoderLayer(nn.Module):
252
self.hidden_size = config.hidden_size
253
max_position_embeddings = getattr(config, "max_position_embeddings", 8192)
254
# Support abacusai/Smaug-72B-v0.1 with attention_bias
255
- # Support internlm/internlm-7b with bias
255
attention_bias = getattr(config, "attention_bias", False) or getattr(
256
config, "bias", False
257
)
vllm/model_executor/models/exaone.py
-1
@@ -243,7 +243,6 @@ class ExaoneDecoderLayer(nn.Module):
243
self.hidden_size = config.hidden_size
244
max_position_embeddings = getattr(config, "max_position_embeddings", 8192)
245
# Support abacusai/Smaug-72B-v0.1 with attention_bias
246
- # Support internlm/internlm-7b with bias
246
attention_bias = getattr(config, "attention_bias", False) or getattr(
247
config, "bias", False
248
)
vllm/model_executor/models/exaone4.py
-1
@@ -230,7 +230,6 @@ class Exaone4DecoderLayer(nn.Module):
230
self.hidden_size = config.hidden_size
231
max_position_embeddings = getattr(config, "max_position_embeddings", 8192)
232
# Support abacusai/Smaug-72B-v0.1 with attention_bias
233
- # Support internlm/internlm-7b with bias
233
attention_bias = getattr(config, "attention_bias", False) or getattr(
234
config, "bias", False
235
)
vllm/model_executor/models/exaone_moe.py
-1
@@ -179,7 +179,6 @@ class ExaoneMoeDecoderLayer(nn.Module):
179
self.hidden_size = config.hidden_size
180
max_position_embeddings = getattr(config, "max_position_embeddings", 8192)
181
# Support abacusai/Smaug-72B-v0.1 with attention_bias
182
- # Support internlm/internlm-7b with bias
182
attention_bias = getattr(config, "attention_bias", False) or getattr(
183
config, "bias", False
184
)
vllm/model_executor/models/granite.py
-1
@@ -199,7 +199,6 @@ class GraniteDecoderLayer(nn.Module):
199
self.residual_multiplier = config.residual_multiplier
200
max_position_embeddings = getattr(config, "max_position_embeddings", 8192)
201
# Support abacusai/Smaug-72B-v0.1 with attention_bias
202
- # Support internlm/internlm-7b with bias
202
attention_bias = getattr(config, "attention_bias", False) or getattr(
203
config, "bias", False
204
)
vllm/model_executor/models/jais2.py
-1
@@ -225,7 +225,6 @@ class Jais2DecoderLayer(nn.Module):
225
self.hidden_size = config.hidden_size
226
max_position_embeddings = getattr(config, "max_position_embeddings", 8192)
227
# Support abacusai/Smaug-72B-v0.1 with attention_bias
228
- # Support internlm/internlm-7b with bias
228
attention_bias = getattr(config, "attention_bias", False) or getattr(
229
config, "bias", False
230
)
vllm/model_executor/models/llama.py
-1
@@ -268,7 +268,6 @@ class LlamaDecoderLayer(nn.Module):
268
self.hidden_size = config.hidden_size
269
max_position_embeddings = getattr(config, "max_position_embeddings", 8192)
270
# Support abacusai/Smaug-72B-v0.1 with attention_bias
271
- # Support internlm/internlm-7b with bias
271
attention_bias = getattr(config, "attention_bias", False) or getattr(
272
config, "bias", False
273
)
vllm/model_executor/models/nemotron.py
-1
@@ -237,7 +237,6 @@ class NemotronDecoderLayer(nn.Module):
237
self.hidden_size = config.hidden_size
238
max_position_embeddings = getattr(config, "max_position_embeddings", 8192)
239
# Support abacusai/Smaug-72B-v0.1 with attention_bias
240
- # Support internlm/internlm-7b with bias
240
attention_bias = getattr(config, "attention_bias", False) or getattr(
241
config, "bias", False
242
)
vllm/model_executor/models/nemotron_nas.py
-1
@@ -141,7 +141,6 @@ class DeciLMDecoderLayer(nn.Module):
141
self.hidden_size = config.hidden_size
142
max_position_embeddings = getattr(config, "max_position_embeddings", 8192)
143
# Support abacusai/Smaug-72B-v0.1 with attention_bias
144
- # Support internlm/internlm-7b with bias
144
attention_bias = getattr(config, "attention_bias", False) or getattr(
145
config, "bias", False
146
)
vllm/model_executor/models/registry.py
+1
-1
@@ -140,7 +140,6 @@ _TEXT_GENERATION_MODELS = {
140
"HCXVisionForCausalLM": ("hyperclovax_vision", "HCXVisionForCausalLM"),
141
"HCXVisionV2ForCausalLM": ("hyperclovax_vision_v2", "HCXVisionV2ForCausalLM"),
142
"HyperCLOVAXForCausalLM": ("hyperclovax", "HyperCLOVAXForCausalLM"),
143
- "InternLMForCausalLM": ("llama", "LlamaForCausalLM"),
143
"InternLM2ForCausalLM": ("internlm2", "InternLM2ForCausalLM"),
144
"InternLM2VEForCausalLM": ("internlm2_ve", "InternLM2VEForCausalLM"),
145
"InternLM3ForCausalLM": ("llama", "LlamaForCausalLM"),
@@ -715,6 +714,7 @@ _PREVIOUSLY_SUPPORTED_MODELS = {
714
"ErnieForTokenClassification": "0.23.0",
715
"QWenLMHeadModel": "0.23.0",
716
"QwenVLForConditionalGeneration": "0.23.0",
717
+ "InternLMForCausalLM": "0.23.0",
718
# encoder-decoder models except whisper
719
# have been removed for V0 deprecation.
720
"DonutForConditionalGeneration": "0.10.2",
vllm/model_executor/models/solar.py
-1
@@ -198,7 +198,6 @@ class SolarDecoderLayer(nn.Module):
198
self.hidden_size = config.hidden_size
199
max_position_embeddings = getattr(config, "max_position_embeddings", 8192)
200
# Support abacusai/Smaug-72B-v0.1 with attention_bias
201
- # Support internlm/internlm-7b with bias
201
attention_bias = getattr(config, "attention_bias", False) or getattr(
202
config, "bias", False
203
)