Upload folder using huggingface_hub (#3)

- ff19bf1b691c6eb4cdebb583aca9d76ca0841296d5c1a633b7eef1526c38bd53 (a967080d12880182048fb425e9773be8d2c8792e)
- 3e40f5f885d6655b42a5706cc3f2cd3567607670a409abaddb2443ca5da6bc5f (822ba84b7a00149b7b387d6510f5c2a65dcee78b)

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/luffycodes-vicuna-class-shishya-7b-ep3-bnb-4bit-smashed",
-                                                 trust_remote_code=True)
+                                                 trust_remote_code=True, device_map='auto')
     tokenizer = AutoTokenizer.from_pretrained("luffycodes/vicuna-class-shishya-7b-ep3")
     
     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/tmpkybk032c",
+  "_name_or_path": "/tmp/tmpgvag0okb",
   "architectures": [
     "LlamaForCausalLM"
   ],
@@ -21,7 +21,7 @@
   "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 sha256:1b8fe45923bee5ba07445b41efb9c95285cfc9c09af250d60ca0f2f2fa9c2656
-size 3866042083
+oid sha256:0a6d3a5c7496c2d078758d334abc7fade42eddb648a499c26ee909e7e39366da
+size 4167722008

smash_config.json

@@ -8,7 +8,7 @@
         "compilers": "None",
         "task": "text_text_generation",
         "device": "cuda",
-        "cache_dir": "/ceph/hdd/staff/charpent/.cache/modelsdwc801x5",
+        "cache_dir": "/ceph/hdd/staff/charpent/.cache/modelssxly_8m6",
         "batch_size": 1,
         "model_name": "luffycodes/vicuna-class-shishya-7b-ep3",
         "pruning_ratio": 0.0,