Upload folder using huggingface_hub

config.json

@@ -12,6 +12,7 @@
   "hidden_size": 8192,
   "initializer_range": 0.02,
   "intermediate_size": 28672,
+  "kv_cache_torch_dtype": "float8_e4m3fn",
   "max_position_embeddings": 4096,
   "mlp_bias": false,
   "model_type": "llama",
@@ -23,7 +24,7 @@
   "rope_scaling": null,
   "rope_theta": 10000.0,
   "tie_word_embeddings": false,
-  "torch_dtype": "float8_e4m3fn",
+  "torch_dtype": "float16",
   "training_pipeline_parallel": 1,
   "training_tensor_parallel": 1,
   "transformers_version": "4.41.0.dev0",

model-00001-of-00029.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1b7a773e6539ff6286f191598136e72fffd6109e87d3587e4f8d569fd3c3cb82
-size 4718660669
+oid sha256:6905b9fd4c8c73b2463d2bb557d7cf4bcc0c2bebdc169abb8590556615f33f55
+size 4718660580

model.safetensors.index.json

@@ -1,11 +1,10 @@
 {
   "metadata": {
-    "total_size": 137953296705
+    "total_size": 137953296704
   },
   "weight_map": {
     "lm_head.weight": "model-00029-of-00029.safetensors",
     "model.embed_tokens.weight": "model-00001-of-00029.safetensors",
-    "model.kv_cache_dtype": "model-00001-of-00029.safetensors",
     "model.layers.0.input_layernorm.weight": "model-00001-of-00029.safetensors",
     "model.layers.0.mlp.down_proj.weight": "model-00001-of-00029.safetensors",
     "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00029.safetensors",