Upload folder using huggingface_hub

config.json

@@ -90,7 +90,7 @@
     "quantization_status": "compressed",
     "sparsity_config": {},
     "transform_config": {},
-    "version": "0.17.1"
+    "version": "0.17.2.a20260730"
   },
   "rms_norm_eps": 1e-05,
   "rope_parameters": {
@@ -98,7 +98,7 @@
     "rope_type": "default"
   },
   "tie_word_embeddings": false,
-  "transformers_version": "5.10.1",
+  "transformers_version": "5.14.1",
   "use_cache": true,
   "vocab_size": 32000
 }

generation_config.json

@@ -3,5 +3,5 @@
   "eos_token_id": 2,
   "max_length": 2048,
   "pad_token_id": 0,
-  "transformers_version": "5.10.1"
+  "transformers_version": "5.14.1"
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c215c008911835e76af967b5073ff860c6bdfef6d49a4cff0b301c8ec96798ba
+oid sha256:ef2040ee3cacd94617be13cd87f1c7ab6216e643dfe064d0c6f36d39dd24315f
 size 1131284624

recipe.yaml

@@ -1,6 +1,7 @@
 quant_stage:
   quant_modifiers:
     AWQModifier:
+      requires_calibration_data: true
       mappings:
       - smooth_layer: re:.*input_layernorm$
         balance_layers: ['re:.*q_proj$', 're:.*k_proj$', 're:.*v_proj$']
@@ -12,11 +13,13 @@ quant_stage:
       ignore: [lm_head]
       scheme: W4A16
       bypass_divisibility_checks: false
+      requires_calibration_data: false
     GPTQModifier:
       targets: ['re:.*mlp\.(down|gate|up)_proj$']
       ignore: [lm_head]
       scheme: W8A8
       bypass_divisibility_checks: false
+      requires_calibration_data: true
       block_size: 128
       dampening_frac: 0.001
       offload_hessians: false