Update space

app.py

@@ -1,4 +1,59 @@
 import gradio as gr
+import torch
+import transformers
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+class CustomPrompter(Prompter):
+ def get_response(self, output: str) -> str:
+ return output.split(self.template["response_split"])[1].strip().split("### Instruction:")[0]
+
+prompter = CustomPrompter(prompt_template_name)
+
+def tokenize(prompt, add_eos_token=True):
+ result = tokenizer(
+ prompt,
+ truncation=True,
+ max_length=cutoff_len,
+ padding=False,
+ return_tensors=None,
+ )
+ if (
+ result["input_ids"][-1] != tokenizer.eos_token_id
+ and len(result["input_ids"]) < cutoff_len
+ and add_eos_token
+ ):
+ result["input_ids"].append(tokenizer.eos_token_id)
+ result["attention_mask"].append(1)
+
+ result["labels"] = result["input_ids"].copy()
+
+ return result
+
+def generate_and_tokenize_prompt(data_point):
+ full_prompt = prompter.generate_prompt(
+ data_point["instruction"],
+ data_point["input"],
+ data_point["output"],
+ )
+ tokenized_full_prompt = tokenize(full_prompt)
+ if not train_on_inputs:
+ user_prompt = prompter.generate_prompt(
+ data_point["instruction"], data_point["input"]
+ )
+ tokenized_user_prompt = tokenize(
+ user_prompt, add_eos_token=add_eos_token
+ )
+ user_prompt_len = len(tokenized_user_prompt["input_ids"])
+
+ if add_eos_token:
+ user_prompt_len -= 1
+
+ tokenized_full_prompt["labels"] = [
+ -100
+ ] * user_prompt_len + tokenized_full_prompt["labels"][
+ user_prompt_len:
+ ] # could be sped up, probably
+ return tokenized_full_prompt
 
 def evaluate(instruction):
  # Generate a response:
@@ -56,6 +111,7 @@ def evaluate(instruction):
  # print(output)
  return(f' {prompter.get_response(output)}')
 
+
 gr.Interface(
  fn=evaluate,
  inputs=[
@@ -73,4 +129,5 @@ gr.Interface(
  ],
  title="🌲 ELM - Erasmian Language Model",
  description="ELM is a 900M parameter language model finetuned to follow instruction. It is trained on Erasmus University academic outputs and the [Stanford Alpaca](https://github.com/tatsu-lab/stanford_alpaca) dataset. For more information, please visit [the GitHub repository](https://github.com/Joaoffg/ELM).", # noqa: E501
- ).queue().load("models/Joaoffg/ELM").launch()
+ ).queue().launch(server_name="0.0.0.0", share=True)
+ # Old testing code follows.