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 15, 09:45 UTC
[CI/Build] Update bamba test model location (#18544)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Harry Mellor committed
May 22, 2025 at 15:01 UTC
ca86a7cf6edbc5baa6a3974d92c2e6eda4fce695
3 files changed
+3
-3
tests/models/language/generation/test_hybrid.py
+1
-1
@@ -31,7 +31,7 @@ HYBRID_MODELS = [
31
# not compatible with pip-compile.
32
"pfnet/plamo-2-1b",
33
"Zyphra/Zamba2-1.2B-instruct",
34
- "hmellor/bamba-tiny-random",
34
+ "hmellor/tiny-random-BambaForCausalLM",
35
]
36
37
# Avoid OOM
tests/models/registry.py
+1
-1
@@ -124,7 +124,7 @@ _TEXT_GENERATION_EXAMPLE_MODELS = {
124
"BaichuanForCausalLM": _HfExamplesInfo("baichuan-inc/Baichuan2-7B-chat",
125
trust_remote_code=True),
126
"BambaForCausalLM": _HfExamplesInfo("ibm-ai-platform/Bamba-9B",
127
- extras={"tiny": "hmellor/bamba-tiny-random"}), # noqa: E501
127
+ extras={"tiny": "hmellor/tiny-random-BambaForCausalLM"}), # noqa: E501
128
"BloomForCausalLM": _HfExamplesInfo("bigscience/bloom-560m",
129
{"1b": "bigscience/bloomz-1b1"}),
130
"ChatGLMModel": _HfExamplesInfo("THUDM/chatglm3-6b",
tests/v1/test_oracle.py
+1
-1
@@ -12,7 +12,7 @@ UNSUPPORTED_MODELS_V1 = [
12
"openai/whisper-large-v3", # transcription
13
"facebook/bart-large-cnn", # encoder decoder
14
"mistralai/Mamba-Codestral-7B-v0.1", # mamba
15
- "hmellor/bamba-tiny-random", # hybrid
15
+ "hmellor/tiny-random-BambaForCausalLM", # hybrid
16
"BAAI/bge-m3", # embedding
17
]
18