justinkarlin
/

idefics-9b-faces

Image-Text-to-Text

text-generation-inference

Inference Endpoints

8-bit precision

bitsandbytes_8bit

Model card Files Files and versions Community

justinkarlin commited on May 1

Commit

e1ea302

•

1 Parent(s): 4907cba

Update config.json

Files changed (1) hide show

config.json +2 -2

config.json CHANGED Viewed

@@ -34,7 +34,7 @@
   "qk_layer_norms": true,
   "quantization_config": {
     "_load_in_4bit": false,
-    "_load_in_8bit": true,
     "bnb_4bit_compute_dtype": "float16",
     "bnb_4bit_quant_storage": "uint8",
     "bnb_4bit_quant_type": "nf4",
@@ -48,7 +48,7 @@
     "llm_int8_threshold": 6.0,
     "load_in_4bit": false,
     "load_in_8bit": false,
-    "quant_method": "bitsandbytes"
   },
   "rms_norm_eps": 1e-06,
   "tie_word_embeddings": false,

   "qk_layer_norms": true,
   "quantization_config": {
     "_load_in_4bit": false,
+    "_load_in_8bit": false,
     "bnb_4bit_compute_dtype": "float16",
     "bnb_4bit_quant_storage": "uint8",
     "bnb_4bit_quant_type": "nf4",
     "llm_int8_threshold": 6.0,
     "load_in_4bit": false,
     "load_in_8bit": false,
+    "quant_method": "bitsandbytes_8bit"
   },
   "rms_norm_eps": 1e-06,
   "tie_word_embeddings": false,