Upload folder using huggingface_hub (#2)

- 8285d1b323d940735cb273b070dd0fb0f3ecdf13f9104c97175f6dc993187651 (161d97f5d9d731a00014e287b31d34c3185ba7f2)
- 9081cbcd5562eac9bd1dfe4cfcc4b5e0cd7edf8555f718ee0d795e6ae733034f (397f3231eea754cda01bad60a9e35d40d717a23a)

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/Vision-CAIR-vicuna-7b-bnb-4bit-smashed",
-                                                 trust_remote_code=True)
+                                                 trust_remote_code=True, device_map='auto')
     tokenizer = AutoTokenizer.from_pretrained("Vision-CAIR/vicuna-7b")
     
     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/tmpgs9760tp",
+  "_name_or_path": "/tmp/tmp25cnfhap",
   "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 sha256:11e99c193bd4518df64bf88ecd657bab5e0464fcc6a34f8ee29ff8fc03889845
-size 3866058445
+oid sha256:dd57a597dc3813743b2289e3a5462a8195e9e4d19019c5be7ff6f36133f535f6
+size 4167738392

smash_config.json

@@ -8,7 +8,7 @@
         "compilers": "None",
         "task": "text_text_generation",
         "device": "cuda",
-        "cache_dir": "/ceph/hdd/staff/charpent/.cache/models6ylrvp34",
+        "cache_dir": "/ceph/hdd/staff/charpent/.cache/modelszbrvxkob",
         "batch_size": 1,
         "model_name": "Vision-CAIR/vicuna-7b",
         "pruning_ratio": 0.0,