From b4fdf7dfe67f7804dfd8c0647e3a4cfd8cc9c0ec Mon Sep 17 00:00:00 2001 From: Flavia Beo Date: Wed, 6 Nov 2024 14:45:35 -0300 Subject: [PATCH] Revert "[Bugfix] Gpt-j-6B patch kv_scale to k_scale path (#10063)" This reverts commit ea928f608c44b825d28609460e0d375a5f877940. --- vllm/model_executor/models/gpt_j.py | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-) diff --git a/vllm/model_executor/models/gpt_j.py b/vllm/model_executor/models/gpt_j.py index b28a6081b868f..76d2729e164e0 100644 --- a/vllm/model_executor/models/gpt_j.py +++ b/vllm/model_executor/models/gpt_j.py @@ -35,8 +35,7 @@ from vllm.model_executor.layers.sampler import Sampler, SamplerOutput from vllm.model_executor.layers.vocab_parallel_embedding import ( ParallelLMHead, VocabParallelEmbedding) -from vllm.model_executor.model_loader.weight_utils import ( - default_weight_loader, maybe_remap_kv_scale_name) +from vllm.model_executor.model_loader.weight_utils import default_weight_loader from vllm.model_executor.sampling_metadata import SamplingMetadata from vllm.sequence import IntermediateTensors @@ -308,9 +307,6 @@ def load_weights(self, weights: Iterable[Tuple[str, torch.Tensor]]): weight_loader(param, loaded_weight, shard_id) break else: - name = maybe_remap_kv_scale_name(name, params_dict) - if name is None: - continue # Skip loading extra bias for GPTQ models. if name.endswith(".bias") and name not in params_dict: continue