Upload folder using huggingface_hub (#2)

- e3b9916fc8f62184454325bb1b525046ef75f52fe6c8e9053b9a1cf5c319a707 (7d6fe6e58190da3aeb560807539c2d0a653bb78a)
- 93ee25177fe4a547193b48ac86187ce7a97b74b0bb49b62c197dfc521af330ed (02af7a275fd01877f256b8ed686c423decf368c9)

README.md

@@ -1,5 +1,4 @@
 ---
-library_name: pruna-engine
 thumbnail: "https://assets-global.website-files.com/646b351987a8d8ce158d1940/64ec9e96b4334c0e1ac41504_Logo%20with%20white%20text.svg"
 metrics:
 - memory_disk
@@ -8,6 +7,8 @@ metrics:
 - inference_throughput
 - inference_CO2_emissions
 - inference_energy_consumption
+tags:
+- pruna-ai
 ---
 <!-- header start -->
 <!-- 200823 -->
@@ -33,7 +34,7 @@ metrics:
 
 ## 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.
@@ -60,7 +61,7 @@ You can run the smashed model with these steps:
    from transformers import AutoModelForCausalLM, AutoTokenizer
 
     model = AutoModelForCausalLM.from_pretrained("PrunaAI/Open-Orca-Mistral-7B-SlimOrca-bnb-4bit-smashed",
-                                                 trust_remote_code=True)
+                                                 trust_remote_code=True, device_map='auto')
     tokenizer = AutoTokenizer.from_pretrained("Open-Orca/Mistral-7B-SlimOrca")
     
     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/tmp5d7b34bh",
+  "_name_or_path": "/tmp/tmphkvf78av",
   "architectures": [
     "MistralForCausalLM"
   ],
@@ -18,7 +18,7 @@
   "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:a57f4c3ffed813582307390a1d747c77f2c4c5a638c92784c3744eb6c45e5c70
-size 4125720389
+oid sha256:5b81de9044ca92278db8315e0f396296145cc678701f82926ef3a618db722cc3
+size 4450870264

smash_config.json

@@ -8,7 +8,7 @@
         "compilers": "None",
         "task": "text_text_generation",
         "device": "cuda",
-        "cache_dir": "/ceph/hdd/staff/charpent/.cache/modelsjl50t8fm",
+        "cache_dir": "/ceph/hdd/staff/charpent/.cache/modelsc_n2zzpc",
         "batch_size": 1,
         "model_name": "Open-Orca/Mistral-7B-SlimOrca",
         "pruning_ratio": 0.0,