Upload folder using huggingface_hub (#3)

- 13d5e7f7ddeba3c064fd58130a143655f9b7a4b0b64a6c671d0daef5ee20e800 (a75cb5995e1b2d7c5db8ab3746220c510fe27fd9)

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/TinyLlama-TinyLlama-1.1B-Chat-v0.3-bnb-4bit-smashed",
-                                                 trust_remote_code=True)
+                                                 trust_remote_code=True, device_map='auto')
     tokenizer = AutoTokenizer.from_pretrained("TinyLlama/TinyLlama-1.1B-Chat-v0.3")
     
     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/tmp92x6w296",
+  "_name_or_path": "/tmp/tmp7rf2uun5",
   "architectures": [
     "LlamaForCausalLM"
   ],
@@ -20,7 +20,7 @@
   "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 sha256:19a45b2138c1b18a29a07322329d391263a45373c097b4239904c47e2e630f3d
-size 762477866
+oid sha256:81f94b76901ae7b56378fe81d598843fd430a03a4a4f75dc3028cfebfa0b4fe9
+size 807450862

smash_config.json

@@ -8,7 +8,7 @@
         "compilers": "None",
         "task": "text_text_generation",
         "device": "cuda",
-        "cache_dir": "/ceph/hdd/staff/charpent/.cache/modelsb4so4lqk",
+        "cache_dir": "/ceph/hdd/staff/charpent/.cache/models_p3vgi1z",
         "batch_size": 1,
         "model_name": "TinyLlama/TinyLlama-1.1B-Chat-v0.3",
         "pruning_ratio": 0.0,