a

app.py

@@ -2,10 +2,31 @@ import gradio as gr
 from huggingface_hub import InferenceClient
 
 """
-For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
+Copied from inference in colab notebook
 """
-client = InferenceClient("llama_lora_model_1")
+from unsloth.chat_templates import get_chat_template
+from unsloth import FastLanguageModel
+
+# IMPORTING MODEL AND TOKENIZER ————————
+
+max_seq_length = 2048 # Choose any! We auto support RoPE Scaling internally!
+dtype = None # None for auto detection. Float16 for Tesla T4, V100, Bfloat16 for Ampere+
+load_in_4bit = True # Use 4bit quantization to reduce memory usage. Can be False.
 
+model, tokenizer = FastLanguageModel.from_pretrained(
+    model_name = "llama_lora_model_1",
+    max_seq_length = max_seq_length,
+    dtype = dtype,
+    load_in_4bit = load_in_4bit,
+)
+
+tokenizer = get_chat_template(
+    tokenizer,
+    chat_template = "llama-3.1",
+)
+FastLanguageModel.for_inference(model) # Enable native 2x faster inference
+
+# RUNNING INFERENCE ————————————————————————
 
 def respond(
     message,
@@ -25,19 +46,56 @@ def respond(
 
     messages.append({"role": "user", "content": message})
 
-    response = ""
-
-    for message in client.chat_completion(
+    inputs = tokenizer.apply_chat_template(
         messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
-    ):
-        token = message.choices[0].delta.content
-
-        response += token
-        yield response
+        tokenize = True,
+        add_generation_prompt = True, # Must add for generation
+        return_tensors = "pt",
+    ).to("cuda")
+
+    outputs = model.generate(input_ids = inputs, max_new_tokens = max_tokens, use_cache = True,
+                            temperature = 1.5, min_p = 0.1)
+    response = tokenizer.batch_decode(outputs)
+
+    yield response
+
+"""
+For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
+"""
+# client = InferenceClient("llama_lora_model_1")
+
+
+# def respond(
+#     message,
+#     history: list[tuple[str, str]],
+#     system_message,
+#     max_tokens,
+#     temperature,
+#     top_p,
+# ):
+#     messages = [{"role": "system", "content": system_message}]
+
+#     for val in history:
+#         if val[0]:
+#             messages.append({"role": "user", "content": val[0]})
+#         if val[1]:
+#             messages.append({"role": "assistant", "content": val[1]})
+
+#     messages.append({"role": "user", "content": message})
+
+#     response = ""
+
+#     for message in client.chat_completion(
+#         messages,
+#         max_tokens=max_tokens,
+#         stream=True,
+#         temperature=temperature,
+#         top_p=top_p,
+#     ):
+#         token = message.choices[0].delta.content
+
+#         response += token
+#         yield response
 
 
 """

requirements.txt

@@ -1 +1,3 @@
-huggingface_hub==0.25.2
+huggingface_hub==0.25.2
+
+unsloth