RAFT adding promptflow flows

raft/azure-ai-studio-ft/pf/raft/extract_final_answer.py

@@ -0,0 +1,11 @@
+from promptflow import tool
+
+# The inputs section will change based on the arguments of the tool function, after you save the code
+# Adding type to arguments and return value will help the system show the types properly
+# Please update the function name/signature per need
+@tool
+def extract_final_answer(cot_answer: str) -> str:
+    """
+    Extracts the final answer from the cot_answer field
+    """
+    return cot_answer.split("<ANSWER>: ")[-1]

raft/azure-ai-studio-ft/pf/raft/flow.dag.yaml

@@ -0,0 +1,46 @@
+id: bring_your_own_data_qna
+name: Bring Your Own Data QnA
+inputs:
+  question:
+    type: string
+    default: How to use SDK V2?
+    is_chat_input: false
+  context:
+    type: string
+    is_chat_input: false
+outputs:
+  answer:
+    type: string
+    reference: ${extract_final_answer.output}
+  explanation:
+    type: string
+    reference: ${llama2_7b_finetuned_model.output}
+nodes:
+- name: construct_prompt_with_context
+  type: prompt
+  source:
+    type: code
+    path: prompt.jinja2
+  inputs:
+    context: ${inputs.context}
+    question: ${inputs.question}
+  use_variants: false
+- name: llama2_7b_finetuned_model
+  type: python
+  source:
+    type: code
+    path: test_llm.py
+  inputs:
+    prompt: ${construct_prompt_with_context.output}
+  use_variants: false
+- name: extract_final_answer
+  type: python
+  source:
+    type: code
+    path: extract_final_answer.py
+  inputs:
+    cot_answer: ${llama2_7b_finetuned_model.output}
+  use_variants: false
+node_variants: {}
+environment:
+  python_requirements_txt: requirements.txt

raft/azure-ai-studio-ft/pf/raft/test_llm.py

@@ -0,0 +1,38 @@
+from promptflow import tool
+from promptflow.connections import CustomConnection
+from openai import OpenAI
+
+
+# The inputs section will change based on the arguments of the tool function, after you save the code
+# Adding type to arguments and return value will help the system show the types properly
+# Please update the function name/signature per need
+@tool
+def classify_intent(prompt: str) -> str:
+
+    endpoint_url = "xxx"
+    api_key = "xxx"
+
+    if not api_key:
+        raise Exception("A key should be provided to invoke the endpoint")
+
+    base_url = endpoint_url + '/v1'
+    client = OpenAI(
+        base_url = base_url,
+        api_key=api_key,
+    )
+
+    deployment_name = "Llama-2-7b-raft-bats-18k-unrrr"
+
+    # COMPLETION API
+    response = client.completions.create(
+        model=deployment_name,
+        prompt=prompt,
+        stop="<STOP>",
+        temperature=0.5,
+        max_tokens=512,
+        top_p=0.1,
+        best_of=1,
+        presence_penalty=0,
+    )
+
+    return response.choices[0].text.strip()

raft/azure-ai-studio-ft/pf/rag/extract_final_answer.py

@@ -0,0 +1,12 @@
+from promptflow import tool
+import re
+
+# The inputs section will change based on the arguments of the tool function, after you save the code
+# Adding type to arguments and return value will help the system show the types properly
+# Please update the function name/signature per need
+@tool
+def extract_final_answer(cot_answer: str) -> str:
+    """
+    Extracts the final answer from the cot_answer field
+    """
+    return re.sub(r'<STOP>\s*$', '', cot_answer.split("<ANSWER>: ")[-1])

raft/azure-ai-studio-ft/pf/rag/flow.dag.yaml

@@ -0,0 +1,46 @@
+id: bring_your_own_data_qna
+name: Bring Your Own Data QnA
+inputs:
+  question:
+    type: string
+    default: How to use SDK V2?
+    is_chat_input: false
+  context:
+    type: string
+    is_chat_input: false
+outputs:
+  answer:
+    type: string
+    reference: ${extract_final_answer.output}
+  explanation:
+    type: string
+    reference: ${llama2_7b_chat_base_model.output}
+nodes:
+- name: llama2_7b_chat_base_model
+  type: custom_llm
+  source:
+    type: package_with_prompt
+    path: llama2_7b_finetuned.jinja2
+    tool: promptflow.tools.open_model_llm.OpenModelLLM.call
+  inputs:
+    api: completion
+    endpoint_name: serverlessEndpoint/Llama-2-7b-chat-gmqyf
+    deployment_name: default
+    temperature: 1
+    max_new_tokens: 500
+    top_p: 1
+    model_kwargs: {}
+    context: ${inputs.context}
+    question: ${inputs.question}
+  use_variants: false
+- name: extract_final_answer
+  type: python
+  source:
+    type: code
+    path: extract_final_answer.py
+  inputs:
+    cot_answer: ${llama2_7b_chat_base_model.output}
+  use_variants: false
+node_variants: {}
+environment:
+  python_requirements_txt: requirements.txt