try gpt4free

Dockerfile

@@ -4,8 +4,12 @@ WORKDIR /code
 
 COPY ./requirements.txt /code/requirements.txt
 
+# For hugging face
 RUN pip install --no-cache-dir --upgrade -r /code/requirements.txt
 
+# For gpt4free
+RUN pip install gpt4free
+
 RUN useradd -m -u 1000 user
 
 USER user

main.py

@@ -5,15 +5,21 @@ Based on: https://huggingface.co/docs/hub/spaces-sdks-docker-first-demo
 
 from fastapi import FastAPI, Request
 
-from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
-from transformers import T5Tokenizer, T5ForConditionalGeneration
+# from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
+# from transformers import T5Tokenizer, T5ForConditionalGeneration
+
+import gpt4free
+from gpt4free import Provider, forefront
 
 
 token_size_limit = None
 
 # FROM: https://huggingface.co/facebook/blenderbot-400M-distill?text=Hey+my+name+is+Thomas%21+How+are+you%3F
-tokenizer = AutoTokenizer.from_pretrained("facebook/blenderbot-400M-distill")
-model = AutoModelForSeq2SeqLM.from_pretrained("facebook/blenderbot-400M-distill")
+
+# LAST USED
+# tokenizer = AutoTokenizer.from_pretrained("facebook/blenderbot-400M-distill")
+# model = AutoModelForSeq2SeqLM.from_pretrained("facebook/blenderbot-400M-distill")
+
 # tokenizer = AutoTokenizer.from_pretrained("facebook/blenderbot-1B-distill")
 # model = AutoModelForSeq2SeqLM.from_pretrained("facebook/blenderbot-1B-distill")
 # token_size_limit = 128
@@ -42,14 +48,20 @@ async def Reply(req: Request):
     msg = request.get('msg')
     print(f'MSG: {msg}')
 
-    input_ids = tokenizer(msg, return_tensors='pt').input_ids  # .to('cuda')
-    output = model.generate(
-        input_ids[:, -token_size_limit:],
-        do_sample=True,
-        temperature=request.get('temperature', 0.9),
-        max_length=request.get('max_length', 100),
-    )
-    reply = tokenizer.batch_decode(output)[0]
+    # Hugging face
+    # input_ids = tokenizer(msg, return_tensors='pt').input_ids  # .to('cuda')
+    # output = model.generate(
+    #     input_ids[:, -token_size_limit:],
+    #     do_sample=True,
+    #     temperature=request.get('temperature', 0.9),
+    #     max_length=request.get('max_length', 100),
+    # )
+    # reply = tokenizer.batch_decode(output)[0]
+
+    # gpt4free
+    # usage theb
+    reply = gpt4free.Completion.create(Provider.Theb, prompt=msg)
+
     print(f'REPLY: {reply}')
     return {'reply': reply}