Update: basic methods and endpoints were added

Dockerfile

@@ -0,0 +1,11 @@
+FROM python:3.10-buster
+
+WORKDIR /app/commandr-api-local
+
+COPY . .
+
+RUN pip3 install --upgrade pip
+
+RUN pip3 install -r requirements.txt
+
+RUN python3 app.py

app.py

@@ -0,0 +1,27 @@
+import uvicorn
+from fastapi import FastAPI
+from pydantic import BaseModel
+from model import get_answer_from_llm
+
+
+class Prompt(BaseModel):
+    prompt: str = ''
+
+
+app = FastAPI(
+    title='CommandRLLMAPI'
+)
+
+
+@app.post("/completion/")
+def get_answer(question: Prompt.prompt):
+    answer = get_answer_from_llm(question)
+    return answer
+
+
+if __name__ == '__main__':
+    uvicorn.run(
+        app,
+        host='0.0.0.0',
+        port=8081
+    )

model.py

@@ -0,0 +1,24 @@
+from transformers import AutoTokenizer, AutoModelForCausalLM
+import torch
+
+device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
+
+model_id = "CohereForAI/c4ai-command-r-v01-4bit"
+tokenizer = AutoTokenizer.from_pretrained(model_id)
+model = AutoModelForCausalLM.from_pretrained(model_id).to(device)
+## <BOS_TOKEN><|START_OF_TURN_TOKEN|><|USER_TOKEN|>Hello, how are you?<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>
+
+
+async def get_answer_from_llm(question: str = None):
+    # Format message with the command-r chat template
+    messages = [{"role": "user", "content": f"{question}"}]
+    input_ids = tokenizer.apply_chat_template(messages, tokenize=True, add_generation_prompt=True, return_tensors="pt")
+    gen_tokens = model.generate(
+        input_ids,
+        max_new_tokens=100,
+        do_sample=True,
+        temperature=0.3,
+        )
+
+    gen_text = await tokenizer.decode(gen_tokens[0])
+    return gen_text

requirements.txt

@@ -0,0 +1,8 @@
+torch
+transformers>=4.39.1
+bitsandbytes
+accelerate
+tokenizers
+pydantic
+fastapi
+uvicorn