LquenS
/

Ovis2-4B-GPTQ-Int4-eager

Image-Text-to-Text

text-generation

4-bit precision

Model card Files Files and versions Community

LquenS commited on 13 days ago

Commit

4daa396

·

verified ·

1 Parent(s): 8813230

Update config.json

flash-attention-2 changed with eager

Files changed (1) hide show

config.json +1 -1

config.json CHANGED Viewed

@@ -9,7 +9,7 @@
   "conversation_formatter_class": "QwenConversationFormatter",
   "disable_tie_weight": false,
   "hidden_size": 2048,
-  "llm_attn_implementation": "flash_attention_2",
   "llm_config": {
     "_attn_implementation_autoset": true,
     "_name_or_path": "Qwen/Qwen2.5-3B-Instruct",

   "conversation_formatter_class": "QwenConversationFormatter",
   "disable_tie_weight": false,
   "hidden_size": 2048,
+  "llm_attn_implementation": "eager",
   "llm_config": {
     "_attn_implementation_autoset": true,
     "_name_or_path": "Qwen/Qwen2.5-3B-Instruct",