SUMIN/HuggingFace_transformer
1
0
Fork 0
Code Issues Pull Requests Actions 7 Packages Projects Releases Wiki Activity

[generate] fix default autocompile case on gpu (#37756)

Browse Source
This commit is contained in:
Joao Gante
2025-04-24 15:08:38 +01:00
committed by GitHub
parent 43bb4c0456
commit 4d64c38593
2 changed files with 3 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
src/transformers/generation/utils.py
View File

@@ -3430,7 +3430,8 @@ class GenerationMixin:
model_kwargs = self._get_initial_cache_position(input_ids, model_kwargs) model_kwargs = self._get_initial_cache_position(input_ids, model_kwargs)
model_forward = self.__call__ model_forward = self.__call__
if self._valid_auto_compile_criteria(model_kwargs, generation_config): compile_forward = self._valid_auto_compile_criteria(model_kwargs, generation_config)
if compile_forward:
os.environ["TOKENIZERS_PARALLELISM"] = "0" os.environ["TOKENIZERS_PARALLELISM"] = "0"
model_forward = self.get_compiled_call(generation_config.compile_config) model_forward = self.get_compiled_call(generation_config.compile_config)

1
src/transformers/modeling_utils.py
View File

@@ -5270,6 +5270,7 @@ class PreTrainedModel(nn.Module, ModuleUtilsMixin, PushToHubMixin, PeftAdapterMi
# Only reset it if not present or different from previous config # Only reset it if not present or different from previous config
if "llama4" in self.config.model_type: # TODO try to enable for FULL COMPILE HYBRID CACHE SUPPORT if "llama4" in self.config.model_type: # TODO try to enable for FULL COMPILE HYBRID CACHE SUPPORT
return self.__call__ return self.__call__
compile_config = compile_config or CompileConfig()
default_config = getattr(self.generation_config, "compile_config", None) or CompileConfig() default_config = getattr(self.generation_config, "compile_config", None) or CompileConfig()
if ( if (
not hasattr(self, "_compiled_call") not hasattr(self, "_compiled_call")