Update app.py

app.py

@@ -1,22 +1,15 @@
 # ADD DISCLAIMERS
-# Q: why is this model so fucking slow? A: because i'm not made of money
-
-import gradio as gr
 import os
 
 os.system('pip install llama-cpp-python transformers torch')
 
+import gradio as gr
 from llama_cpp import Llama
 from transformers import AutoTokenizer
-import torch
 from huggingface_hub import upload_file 
 import json
 from uuid import uuid4
 
-# For logging
-def upload_json_to_hub(dict, file_id):
-    upload_file(path_or_fileobj=json.dumps(dict).encode('utf-8'), path_in_repo=file_id, repo_id="Elijahbodden/EliGPT-convologs", token="os.getenv['HF_API_TOKEN']", repo_type="dataset")
-
 model_id = "Elijahbodden/eliGPTv1.1"
 
 # MODEL
@@ -66,6 +59,9 @@ presets = {
     "newcomer" : [{"role": "user", "content": "Hi, how are you?\n"}, {"role": "assistant", "content": "Hiiii!\n I don't think we've ever talked before, nice to meet you\n"}],
 }
 
+# For logging
+def upload_json_to_hub(dict, file_id):
+    upload_file(path_or_fileobj=json.dumps(dict).encode('utf-8'), path_in_repo=file_id, repo_id="Elijahbodden/EliGPT-convologs", token="os.getenv['HF_API_TOKEN']", repo_type="dataset")
 
 def custom_lp_logits_processor(ids, logits, lp_start, lp_decay, prompt_tok_len):
     generated_tok_number = len(ids) - prompt_tok_len
@@ -100,7 +96,6 @@ def respond(
     response = ""
 
     convo = tokenizer.apply_chat_template(messages, tokenize=True)
-    # print(convo)
     for message in model.create_completion(
         convo,
         temperature=0.75,