Upload folder using huggingface_hub

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/bigcode-santacoder-bnb-4bit-smashed",
-                                                 trust_remote_code=True)
+                                                 trust_remote_code=True, device_map='auto')
     tokenizer = AutoTokenizer.from_pretrained("bigcode/santacoder")
     
     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/tmppkofl_7p",
+  "_name_or_path": "/tmp/tmpa5unfhvu",
   "activation_function": "gelu_fast",
   "architectures": [
     "GPT2LMHeadCustomModel"
@@ -24,7 +24,7 @@
   "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 sha256:2dd5eb29012210f57e6646cfe88d931ee3cf621e4d652dd7a14e849be866b437
-size 838045874
+oid sha256:e6eec412e98653f659d74a1b82f125d1bdf8748813e514a10a193b62cd34e539
+size 885412960

smash_config.json

@@ -8,7 +8,7 @@
         "compilers": "None",
         "task": "text_text_generation",
         "device": "cuda",
-        "cache_dir": "/ceph/hdd/staff/charpent/.cache/modelsed4p5bm5",
+        "cache_dir": "/ceph/hdd/staff/charpent/.cache/modelst1744evx",
         "batch_size": 1,
         "model_name": "bigcode/santacoder",
         "pruning_ratio": 0.0,