Upload folder using huggingface_hub (#2)

- c85bba27e7d88d236ebc2090cdb6d49413aec91af27d4c60d2306f73e8cf28d2 (f324a3ac54fea51e20753c51eb58452ebdc9410e)
- f464c7ac68a904ce7a674095c5b5472d04360c1f5f5cb28ce6dcc2ddbbf583cb (3612c1f23e4a7c2b5838dd9ea2c7de05b130e92b)

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/WizardLM-WizardCoder-Python-7B-V1.0-bnb-4bit-smashed",
-                                                 trust_remote_code=True)
+                                                 trust_remote_code=True, device_map='auto')
     tokenizer = AutoTokenizer.from_pretrained("WizardLM/WizardCoder-Python-7B-V1.0")
     
     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/tmpopo2l975",
+  "_name_or_path": "/tmp/tmpbz2lo5f5",
   "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:dee713b1e79cdd9336c0436fa0a76ae7277e3580e84932b15ca899387225fd5a
-size 3866058433
+oid sha256:a7ac073d218fbfaef7da78d8c1231ad41f7a70d2d3104392036db581a80059c5
+size 4167738392

smash_config.json

@@ -8,7 +8,7 @@
         "compilers": "None",
         "task": "text_text_generation",
         "device": "cuda",
-        "cache_dir": "/ceph/hdd/staff/charpent/.cache/modelss1ryfbjq",
+        "cache_dir": "/ceph/hdd/staff/charpent/.cache/modelsqyfn57zp",
         "batch_size": 1,
         "model_name": "WizardLM/WizardCoder-Python-7B-V1.0",
         "pruning_ratio": 0.0,