From 75952729eb90eb7f7da5ba76416516ac8e5bf1c8 Mon Sep 17 00:00:00 2001 From: smthemex <138738845+smthemex@users.noreply.github.com> Date: Mon, 30 Mar 2026 08:35:04 +0800 Subject: [PATCH] Update t5_gemma_model.py --- inference/model/t5_gemma/t5_gemma_model.py | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/inference/model/t5_gemma/t5_gemma_model.py b/inference/model/t5_gemma/t5_gemma_model.py index 14a25f7..da4ad0c 100644 --- a/inference/model/t5_gemma/t5_gemma_model.py +++ b/inference/model/t5_gemma/t5_gemma_model.py @@ -24,7 +24,7 @@ class T5GemmaEncoder: self.gguf_mode=False if model_path is not None: if os.path.isfile(model_path): - configs=T5GemmaConfig.from_pretrained(repo,is_encoder_decoder=False,model_type=weight_dtype,) + configs=T5GemmaConfig.from_pretrained(repo,is_encoder_decoder=False,model_type=weight_dtype,ignore_mismatched_sizes=True,low_cpu_mem_usage=False) with ctx(): model = T5GemmaEncoderModel(configs) model_dict=load_file(model_path) @@ -43,7 +43,7 @@ class T5GemmaEncoder: ) self.model = CPUOffloadWrapper(model, is_cpu_offload=env_is_true("CPU_OFFLOAD") or get_arch_memory() <= 48) elif gguf_path is not None: - configs=T5GemmaConfig.from_pretrained(repo,is_encoder_decoder=False,model_type=weight_dtype,) + configs=T5GemmaConfig.from_pretrained(repo,is_encoder_decoder=False,model_type=weight_dtype,ignore_mismatched_sizes=True,low_cpu_mem_usage=False) with ctx(): self.model = T5GemmaEncoderModel(configs) g_dict=load_gguf_checkpoint(gguf_path) @@ -159,4 +159,4 @@ def set_gguf2meta_model(meta_model,model_state_dict,dtype,device): del model_state_dict gc.collect() - return meta_model.to(dtype=dtype) \ No newline at end of file + return meta_model.to(dtype=dtype)