yzsydlc
/

qwen-primeai-gptq-wiki-4bit

Text Generation

4-bit precision

Model card Files Files and versions Community

yzsydlc commited on Mar 5

Commit

e29461e

•

1 Parent(s): f00275b

add file

Files changed (2) hide show

config.json +1 -1
modeling_qwen.py +1 -1

config.json CHANGED Viewed

@@ -10,7 +10,7 @@
   },
   "bf16": false,
   "emb_dropout_prob": 0.0,
-  "fp16": false,
   "fp32": false,
   "hidden_size": 4096,
   "initializer_range": 0.02,

   },
   "bf16": false,
   "emb_dropout_prob": 0.0,
+  "fp16": true,
   "fp32": false,
   "hidden_size": 4096,
   "initializer_range": 0.02,

modeling_qwen.py CHANGED Viewed

@@ -1057,7 +1057,7 @@ class QWenLMHeadModel(QWenPreTrainedModel):
         )
         hidden_states = transformer_outputs[0]
-        hidden_states = hidden_states.to(torch.bfloat16)
         #解决self和mat2数据类型不一致的bug

         )
         hidden_states = transformer_outputs[0]
+        #hidden_states = hidden_states.to(torch.float16)
         #解决self和mat2数据类型不一致的bug