Upload folder using huggingface_hub (#3)

- 2705e5b69e1bd80a09ac8c5ae81313c4a099274a7c2ab290883fdc61147c1dd9 (e040b5d1d2b3ae577f863ce9aa22d1f8533d8852)
- c293b8592ec73aac9ddf01c88efa72bdbee9d098d144e245bca81f59044bf63e (32d3a2e1352dfc3dcd0bbb565557aaa3fdd3144c)

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/GritLM-GritLM-7B-bnb-4bit-smashed",
-                                                 trust_remote_code=True)
+                                                 trust_remote_code=True, device_map='auto')
     tokenizer = AutoTokenizer.from_pretrained("GritLM/GritLM-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/tmps168pfcr",
+  "_name_or_path": "/tmp/tmptoahj57j",
   "architectures": [
     "MistralForCausalLM"
   ],
@@ -29,7 +29,7 @@
   "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:a674cb4bb36dde94f1487bf15187bc75a007263f44090c66d9f409e068db231b
-size 4125687624
+oid sha256:4961a6835a91120b3ba0de7dc58bf038061d6aefd224a4f383213ca4035821ae
+size 4450837496

smash_config.json

@@ -8,7 +8,7 @@
         "compilers": "None",
         "task": "text_text_generation",
         "device": "cuda",
-        "cache_dir": "/ceph/hdd/staff/charpent/.cache/models3w8ei3_b",
+        "cache_dir": "/ceph/hdd/staff/charpent/.cache/modelspellva0l",
         "batch_size": 1,
         "model_name": "GritLM/GritLM-7B",
         "pruning_ratio": 0.0,