Upload folder using huggingface_hub (#3)

- b0ef2e7a4aca1620ecaf95b6b04306b83eea4c0bc7c34604cae0de4647951019 (5fe0131089c9fd5591c30827e85827714a6c2dec)

README.md

@@ -34,7 +34,7 @@ tags:
 
 ## Results
 
-Detailed efficiency metrics coming soon!
+![image info](./plots.png)
 
 **Frequently Asked Questions**
 - ***How does the compression work?*** The model is compressed with llm-int8.
@@ -61,7 +61,7 @@ You can run the smashed model with these steps:
    from transformers import AutoModelForCausalLM, AutoTokenizer
 
     model = AutoModelForCausalLM.from_pretrained("PrunaAI/ahxt-LiteLlama-460M-1T-bnb-4bit-smashed",
-                                                 trust_remote_code=True)
+                                                 trust_remote_code=True, device_map='auto')
     tokenizer = AutoTokenizer.from_pretrained("ahxt/LiteLlama-460M-1T")
     
     input_ids = tokenizer("What is the color of prunes?,", return_tensors='pt').to(model.device)["input_ids"]

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "/tmp/tmp6y9ocq4m",
+  "_name_or_path": "/tmp/tmpei65drfe",
   "architectures": [
     "LlamaForCausalLM"
   ],
@@ -22,7 +22,7 @@
   "quantization_config": {
     "bnb_4bit_compute_dtype": "bfloat16",
     "bnb_4bit_quant_type": "fp4",
-    "bnb_4bit_use_double_quant": true,
+    "bnb_4bit_use_double_quant": false,
     "llm_int8_enable_fp32_cpu_offload": false,
     "llm_int8_has_fp16_weight": false,
     "llm_int8_skip_modules": [

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:31cb37333b06697763879fb0e9a5ca5a4d383715a4042f2297dfc0fffca6ff28
-size 391477950
+oid sha256:7458c4e7926d62d4d16094c53ef2f85bd7a83df86fd4d448a7dac4d4caba5389
+size 407974704

smash_config.json

@@ -8,7 +8,7 @@
         "compilers": "None",
         "task": "text_text_generation",
         "device": "cuda",
-        "cache_dir": "/ceph/hdd/staff/charpent/.cache/modelstjvx_ujw",
+        "cache_dir": "/ceph/hdd/staff/charpent/.cache/models7pnido7h",
         "batch_size": 1,
         "model_name": "ahxt/LiteLlama-460M-1T",
         "pruning_ratio": 0.0,