Upload folder using huggingface_hub (#5)

- 5772fce89ebdf607c820cc9d25739cb0d2630cd2f138f45553449b5456d2807d (02982c251d5d896a67bdb96b4c3caa97894fdcce)

README.md

@@ -60,8 +60,10 @@ You can run the smashed model with these steps:
     ```python 
    from transformers import AutoModelForCausalLM, AutoTokenizer
 
-    model = AutoModelForCausalLM.from_pretrained("PrunaAI/facebook-opt-125m-HQQ-1bit-smashed",
-                                                 trust_remote_code=True, device_map='auto')
+    try: 
+    model = HQQModelForCausalLM.from_quantized("PrunaAI/facebook-opt-125m-HQQ-1bit-smashed", device_map='auto')
+except: 
+    model = AutoHQQHFModel.from_quantized("PrunaAI/facebook-opt-125m-HQQ-1bit-smashed")
     tokenizer = AutoTokenizer.from_pretrained("facebook/opt-125m")
     
     input_ids = tokenizer("What is the color of prunes?,", return_tensors='pt').to(model.device)["input_ids"]

smash_config.json

@@ -14,7 +14,7 @@
         "controlnet": "None",
         "unet_dim": 4,
         "device": "cuda",
-        "cache_dir": "/ceph/hdd/staff/charpent/.cache/models__u_iaf9",
+        "cache_dir": "/ceph/hdd/staff/charpent/.cache/modelsh3ybd3wx",
         "batch_size": 1,
         "tokenizer": "GPT2TokenizerFast(name_or_path='facebook/opt-125m', vocab_size=50265, model_max_length=1000000000000000019884624838656, is_fast=True, padding_side='right', truncation_side='right', special_tokens={'bos_token': '</s>', 'eos_token': '</s>', 'unk_token': '</s>', 'pad_token': '<pad>'}, clean_up_tokenization_spaces=True),  added_tokens_decoder={\n\t1: AddedToken(\"<pad>\", rstrip=False, lstrip=False, single_word=False, normalized=True, special=True),\n\t2: AddedToken(\"</s>\", rstrip=False, lstrip=False, single_word=False, normalized=True, special=True),\n}",
         "task": "text_text_generation",