Update app.py

app.py

@@ -10,29 +10,29 @@ from transformers import (
 )
 from threading import Thread
 
-## The huggingface model id for Microsoft's phi-2 model
-#checkpoint = "microsoft/phi-2"
+# The huggingface model id for Microsoft's phi-2 model
+checkpoint = "microsoft/phi-2"
 
-## Download and load model and tokenizer
-#tokenizer = AutoTokenizer.from_pretrained(checkpoint, trust_remote_code=True)
-#model = AutoModelForCausalLM.from_pretrained(
-#    checkpoint, torch_dtype=torch.float32, device_map="cpu", trust_remote_code=True
-#)
+# Download and load model and tokenizer
+tokenizer = AutoTokenizer.from_pretrained(checkpoint, trust_remote_code=True)
+model = AutoModelForCausalLM.from_pretrained(
+    checkpoint, torch_dtype=torch.float32, device_map="cpu", trust_remote_code=True
+)
 
-model_name_or_path = "TheBloke/phi-2-GPTQ"
-# To use a different branch, change revision
-# For example: revision="gptq-4bit-32g-actorder_True"
+#model_name_or_path = "TheBloke/phi-2-GPTQ"
+## To use a different branch, change revision
+## For example: revision="gptq-4bit-32g-actorder_True"
 
-config = AutoConfig.from_pretrained(model_name_or_path,trust_remote_code=True)
-config.quantization_config["use_exllama"] = False
+#config = AutoConfig.from_pretrained(model_name_or_path,trust_remote_code=True)
+#config.quantization_config["use_exllama"] = False
 
-model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
-                                             device_map="cpu",
-                                             trust_remote_code=True,
-                                             revision="main",
-                                             config=config)
+#model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
+#                                             device_map="cpu",
+#                                             trust_remote_code=True,
+#                                             revision="main",
+#                                             config=config)
 
-tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
+#tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
 
 
 # Text generation pipeline