Update Space

app.py

@@ -2,21 +2,14 @@ import gradio as gr
 from transformers import AutoTokenizer, AutoModelForCausalLM
 import torch
 
-# Load a small DeepSeek Coder model suitable for CPU and limited RAM usage
-model_name = "deepseek-ai/deepseek-coder-1.3b-base"  # Change to smaller model for your RAM if needed
+model_name = "deepseek-ai/deepseek-coder-1.3b-base"
 
-# Load tokenizer and model
 tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
 model = AutoModelForCausalLM.from_pretrained(model_name, trust_remote_code=True)
-
-# Put model in eval mode (no training)
 model.eval()
 
-def generate_code(prompt, max_tokens, temperature, top_p):
-    # Tokenize input prompt
+def respond(prompt, max_tokens, temperature, top_p):
     inputs = tokenizer(prompt, return_tensors="pt")
-
-    # Generate output tokens
     outputs = model.generate(
         **inputs,
         max_new_tokens=max_tokens,
@@ -25,28 +18,17 @@ def generate_code(prompt, max_tokens, temperature, top_p):
         do_sample=True,
         pad_token_id=tokenizer.eos_token_id,
     )
+    return tokenizer.decode(outputs[0], skip_special_tokens=True)[len(prompt):].strip()
 
-    # Decode generated tokens to string
-    generated_text = tokenizer.decode(outputs[0], skip_special_tokens=True)
-
-    # Return generated completion excluding the input prompt for clarity
-    return generated_text[len(prompt):].strip()
-
-# Gradio app interface
 with gr.Blocks() as demo:
-    gr.Markdown("# DeepSeek Coder Chatbot")
-    prompt_input = gr.Textbox(label="Code Prompt", lines=5, placeholder="Write your code prompt here...")
-    max_tokens_slider = gr.Slider(1, 1024, value=512, step=1, label="Max Generated Tokens")
-    temperature_slider = gr.Slider(0.1, 1.0, value=0.7, step=0.05, label="Temperature")
-    top_p_slider = gr.Slider(0.1, 1.0, value=0.9, step=0.05, label="Top-p (nucleus sampling)")
-    generate_btn = gr.Button("Generate Code")
-    output = gr.Textbox(label="Generated Code", lines=15)
-
-    generate_btn.click(
-        fn=generate_code,
-        inputs=[prompt_input, max_tokens_slider, temperature_slider, top_p_slider],
-        outputs=output,
-    )
+    gr.Markdown("# DeepSeek Coder without Login")
+    prompt = gr.Textbox(label="Enter your prompt", lines=5)
+    max_tokens = gr.Slider(1, 1024, value=512, step=1, label="Max Tokens")
+    temperature = gr.Slider(0.1, 1.0, value=0.7, step=0.05, label="Temperature")
+    top_p = gr.Slider(0.1, 1.0, value=0.9, step=0.05, label="Top-p")
+    btn = gr.Button("Generate")
+    output = gr.Textbox(label="Output", lines=15)
+    btn.click(respond, inputs=[prompt, max_tokens, temperature, top_p], outputs=output)
 
 if __name__ == "__main__":
     demo.launch()