Update app.py

app.py

@@ -42,21 +42,26 @@ import torch
 #         load_in_4bit = True,
 #     )
 # tokenizer = AutoTokenizer.from_pretrained("FlawedLLM/Bhashini")
-# Load model directly
-from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig, AutoConfig
+# # Load model directly
+# from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig, AutoConfig
 
-tokenizer = AutoTokenizer.from_pretrained("FlawedLLM/Bhashini_9")
-config = AutoConfig.from_pretrained("FlawedLLM/Bhashini_9")  # Load configuration
+# tokenizer = AutoTokenizer.from_pretrained("FlawedLLM/Bhashini_9")
+# config = AutoConfig.from_pretrained("FlawedLLM/Bhashini_9")  # Load configuration
 
-# quantization_config = BitsAndBytesConfig(
-#         load_in_4bit=True,
-#         bnb_4bit_use_double_quant=True,
-#         bnb_4bit_quant_type="nf4",  
-#         bnb_4bit_compute_dtype=torch.float16
-# )
+# # quantization_config = BitsAndBytesConfig(
+# #         load_in_4bit=True,
+# #         bnb_4bit_use_double_quant=True,
+# #         bnb_4bit_quant_type="nf4",  
+# #         bnb_4bit_compute_dtype=torch.float16
+# # )
+
+# # torch_dtype =torch.float16
+# model = AutoModelForCausalLM.from_pretrained("FlawedLLM/Bhashini_9",config=config, ignore_mismatched_sizes=True).to('cuda')
+# Load model directly
+from transformers import AutoTokenizer, AutoModelForCausalLM
 
-# torch_dtype =torch.float16
-model = AutoModelForCausalLM.from_pretrained("FlawedLLM/Bhashini_9",config=config, ignore_mismatched_sizes=True).to('cuda')
+tokenizer = AutoTokenizer.from_pretrained("FlawedLLM/Bhashini_00")
+model = AutoModelForCausalLM.from_pretrained("FlawedLLM/Bhashini_00").to('cuda')
 
 @spaces.GPU(duration=300)
 def chunk_it(input_command):