A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 14, 05:30 UTC
[CI Failure] Fix tests with missing TinyLlama-1.1B-Chat-v1.0-FP8-e2e (#26816)
Signed-off-by: mgoin <mgoin64@gmail.com>
Michael Goin committed
Oct 14, 2025 at 14:38 UTC
c3a722fcb2c95c61af3ce5e822b6eb8285abeb85
3 files changed
+3
-3
tests/compile/test_async_tp.py
+1
-1
@@ -332,7 +332,7 @@ def async_tp_pass_on_test_model(
332
333
# this is a fake model name to construct the model config
334
# in the vllm_config, it's not really used.
335
- model_name = "nm-testing/TinyLlama-1.1B-Chat-v1.0-FP8-e2e"
335
+ model_name = "RedHatAI/Llama-3.2-1B-Instruct-FP8"
336
vllm_config.model_config = ModelConfig(
337
model=model_name, trust_remote_code=True, dtype=dtype, seed=42
338
)
tests/compile/test_fusion_all_reduce.py
+1
-1
@@ -229,7 +229,7 @@ def all_reduce_fusion_pass_on_test_model(
229
230
# this is a fake model name to construct the model config
231
# in the vllm_config, it's not really used.
232
- model_name = "nm-testing/TinyLlama-1.1B-Chat-v1.0-FP8-e2e"
232
+ model_name = "RedHatAI/Llama-3.2-1B-Instruct-FP8"
233
vllm_config.model_config = ModelConfig(
234
model=model_name, trust_remote_code=True, dtype=dtype, seed=42
235
)
tests/compile/test_sequence_parallelism.py
+1
-1
@@ -278,7 +278,7 @@ def sequence_parallelism_pass_on_test_model(
278
279
# this is a fake model name to construct the model config
280
# in the vllm_config, it's not really used.
281
- model_name = "nm-testing/TinyLlama-1.1B-Chat-v1.0-FP8-e2e"
281
+ model_name = "RedHatAI/Llama-3.2-1B-Instruct-FP8"
282
vllm_config.model_config = ModelConfig(
283
model=model_name, trust_remote_code=True, dtype=dtype, seed=42
284
)