Add the support for the qwen3 next model (a hybrid attention model). (#24526)

Signed-off-by: Tao He <linzhu.ht@alibaba-inc.com>
Co-authored-by: Jee Jee Li <pandaleefree@gmail.com>
This commit is contained in:
Tao He
2025-09-11 15:32:09 +08:00
committed by GitHub
co-authored by Jee Jee Li
parent 2048c4e379
commit e93f4cc9e3
29 changed files with 2476 additions and 61 deletions
+5 -1
View File
@@ -326,6 +326,8 @@ _TEXT_GENERATION_EXAMPLE_MODELS = {
"Qwen2MoeForCausalLM": _HfExamplesInfo("Qwen/Qwen1.5-MoE-A2.7B-Chat"),
"Qwen3ForCausalLM": _HfExamplesInfo("Qwen/Qwen3-8B"),
"Qwen3MoeForCausalLM": _HfExamplesInfo("Qwen/Qwen3-30B-A3B"),
"Qwen3NextForCausalLM": _HfExamplesInfo("Qwen/Qwen3-Next-80B-A3B-Instruct",
min_transformers_version="4.56.2"),
"RWForCausalLM": _HfExamplesInfo("tiiuae/falcon-40b"),
"SeedOssForCausalLM": _HfExamplesInfo("ByteDance-Seed/Seed-OSS-36B-Instruct", # noqa: E501
trust_remote_code=True,
@@ -640,7 +642,9 @@ _SPECULATIVE_DECODING_EXAMPLE_MODELS = {
is_available_online=False),
"MiMoMTPModel": _HfExamplesInfo("XiaomiMiMo/MiMo-7B-RL",
trust_remote_code=True,
speculative_model="XiaomiMiMo/MiMo-7B-RL")
speculative_model="XiaomiMiMo/MiMo-7B-RL"),
"Qwen3NextMTP": _HfExamplesInfo("Qwen/Qwen3-Next-80B-A3B-Instruct",
min_transformers_version="4.56.2"),
}
_TRANSFORMERS_BACKEND_MODELS = {