Upload folder using huggingface_hub

README.md

@@ -76,7 +76,7 @@ with open(hf_hub_download(source_model_id, filename='config.json', repo_type='mo
     config_json = json.load(f)
 config_json.update({
     "head_dim": 32,
-    "hidden_size": 16,
+    "hidden_size": 32,  # required by Mxfp4GptOssExperts codes
     "intermediate_size": 64,
     "layer_types": ["sliding_attention", "full_attention"],
     "num_attention_heads": 2,
@@ -106,6 +106,7 @@ with torch.no_grad():
 model.save_pretrained(save_folder)
 
 # mxfp4
+from transformers.quantizers.quantizer_mxfp4 import Mxfp4HfQuantizer
 # model = AutoModelForCausalLM.from_pretrained(save_folder, trust_remote_code=True, torch_dtype=torch.bfloat16, quantization_config=quantization_config)
 # model.save_pretrained(save_folder, safe_serialization=True)
 ```

config.json

@@ -8,7 +8,7 @@
   "experts_per_token": 4,
   "head_dim": 32,
   "hidden_act": "silu",
-  "hidden_size": 16,
+  "hidden_size": 32,
   "initial_context_length": 4096,
   "initializer_range": 0.02,
   "intermediate_size": 64,

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:304de707ed9cbe5bc3a7e574bb7da85f7afcbb9733cb39ada46200a17b740308
-size 6865464
+oid sha256:aefe8b9c4b4969f6d13c5d778760f3dce4e25134324b33677934550d9df02a7c
+size 13710176