token

app.py

@@ -13,28 +13,23 @@ huggingface_token = os.getenv('HUGGINGFACE_TOKEN')
 if not huggingface_token:
     raise Exception("Hugging Face token not found. Please set it as an environment variable 'HUGGINGFACE_TOKEN'.")
 
-# Define cache directory
-cache_dir = "./cache"
-
 # Load the base model without quantization configuration
 base_model = AutoModelForCausalLM.from_pretrained(
     base_model_id,  
     trust_remote_code=True,
-    token=huggingface_token,  # Use the token parameter
-    cache_dir=cache_dir  # Specify cache directory
-)
+    token=huggingface_token  # Use the token parameter
+).to("cuda")  # Move model to CUDA
 
 # Load the tokenizer
 tokenizer = AutoTokenizer.from_pretrained(
     base_model_id, 
     add_bos_token=True, 
     trust_remote_code=True, 
-    token=huggingface_token,
-    cache_dir=cache_dir  # Specify cache directory
+    token=huggingface_token
 )
 
-# Load the fine-tuned model
-ft_model = PeftModel.from_pretrained(base_model, "checkpoint-2800", cache_dir=cache_dir)
+# Load the fine-tuned model and move to CUDA
+ft_model = PeftModel.from_pretrained(base_model, "checkpoint-2800").to("cuda")  # Move model to CUDA
 
 def formatting_func(job_description):
     text = f"### The job description: {job_description}\n ### The skills: "
@@ -66,4 +61,4 @@ inputs = gr.Textbox(lines=10, label="Job description:", placeholder="Enter or pa
 outputs = gr.Textbox(label="Required skills:", placeholder="The required skills will be displayed here...")
 
 gr.Interface(fn=generate_skills, inputs=inputs, outputs=outputs, title="Job Skills Analysis",
-             description="Paste the job description in the text box below and the model will show the required skills for candidates.").launch()  # Remove share=True
+             description="Paste the job description in the text box below and the model will show the required skills for candidates.").launch()