Upload folder using huggingface_hub (#3)

- dc63db36c355185fe2e9029ccfa0aea8aeca91088823c8494c46b1b17ae98909 (21c4f1072c18012cbecb9938d0b31f7c667ac332)
- 4731e2c1c27764403cb5fa44dc2bf7b89c3de00be3e2f10a2a28f1915408a853 (c680b4fd4f7dccd72c1f3a783c5e6af8694f3b39)

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/EleutherAI-llemma_7b-bnb-4bit-smashed",
-                                                 trust_remote_code=True)
+                                                 trust_remote_code=True, device_map='auto')
     tokenizer = AutoTokenizer.from_pretrained("EleutherAI/llemma_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/tmphn7syz0z",
+  "_name_or_path": "/tmp/tmpkfcz2zz8",
   "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 sha256:75dab0f182f72f9d522afcec6141b68d34bb54f12d4f3155199a340a44128e58
-size 3866304196
+oid sha256:03faf0049003de2934362b9392ec14a5d790df26e70d8c00ddf6cbe0085a8559
+size 4167984152

smash_config.json

@@ -8,7 +8,7 @@
         "compilers": "None",
         "task": "text_text_generation",
         "device": "cuda",
-        "cache_dir": "/ceph/hdd/staff/charpent/.cache/modelsljtsm849",
+        "cache_dir": "/ceph/hdd/staff/charpent/.cache/models50yli0mx",
         "batch_size": 1,
         "model_name": "EleutherAI/llemma_7b",
         "pruning_ratio": 0.0,