[Model] Use upstream GLM-5.3 and Qwen4-Exp configs and processor (#57387)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com> Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
This commit is contained in:
co-authored by
Claude Opus 5
parent
866f02db8a
commit
e3cae8d2ac
@@ -1059,7 +1059,7 @@ class SpeculativeConfig:
|
||||
)
|
||||
if hf_config.model_type == "glm5_next":
|
||||
hf_config.model_type = "glm5_next_mtp"
|
||||
n_predict = hf_config.num_nextn_predict_layers
|
||||
n_predict = hf_config.get_text_config().num_nextn_predict_layers
|
||||
hf_config.update(
|
||||
{"n_predict": n_predict, "architectures": ["Glm5NextMTPModel"]}
|
||||
)
|
||||
|
||||
Reference in New Issue
Block a user