Python Code Needs to be Converted into GPT Actions that are located at OPENAI - 23/01/2024 23:51 EST
- Tila: Closed
- Palkinto: $40
- Vastaanotetut työt: 5
- Voittaja: achyuttiwari22
Kilpailun tehtävänanto
I'm looking for a skilled coder who can transform our existing Python code(attached below) into a schema and GPT's actions. To be Clear this code uses the OPEN_AI API. However, it does not require a API Key because it is a GPTs that is hosted by the main AI company "OPEN AI" at https://chat.openai.com/ SEE CODE ATTACHED below that needs to be converted. This transformation aims to enable all processing and functions to be performed using the GPT and its Actions feature, I think any Endpoint or API action can be utilized and the assistants can have the ability to call tools as well. I think OPEN Ai takes my instructions in the first text box as input then like the example of a web apis it delivers client requests and returns responses via JSON. Each Request response cycle is an "API call" The API call consists of a server endpoint URL and a request method. Usually through HTTP. So this is area of detrmining the coding that I need help with. For example, "GET/weather HTTP/2" You probably need to look at the so called Cookbook published by OPEN AI in order to know how to take the responses you get from a call to the ChatGPT model and then https://cookbook.openai.com/examples/how_to_call_functions_with_chat_models the GPTs Specs to understand how to do it. You may have to implement forcing it to execute a function as it could lose attention or hallucinate. Use "# in this cell we force the model to use get_n_day_weather_forecast
messages = []
messages.append({"role": "system", "content": "Don't make assumptions about what values to plug into functions. Ask for clarification if a user request is ambiguous."})
messages.append({"role": "user", "content": "Give me a weather report for Toronto, Canada."})
chat_response = chat_completion_request(
messages, tools=tools, tool_choice={"type": "function", "function": {"name": "get_n_day_weather_forecast"}}
)
chat_response.json()["choices"][0]["message"]" Then you may have to include schema instruction on how to execute functions whose inputs are model-generated, and use this to implement an agent that can answer questions for us about the output. In the functionality we need for this project the ai will 1. Generate 3 "working theories" on which is the best response to provide the user. That particular output should have three labeled options, an explorations and description of each of the three options. The last output then becomes then input for the grading and ranking step. The input can be the previous output with a prompt for the agent to answer questions about the number rank and criteria grades for the immediate prior output. Then it gets complex but the lowest two are ignored and the Agent is asked to Explore the winning idea maybe include an executive summary. The next step the Agent is asked to generate two more competing ideas. Then a loop occurs because we are repeating the print all three descriptions of the 3 competing ideas. The repeat of the step involving grading also occurs so that you end up with ranked output and scores for each of the proposed solution to the users problem. This looping occurs as programmed a minimum of 5 times. After the loop is complete you have one winning output.
The final step is printing an Executive summary of the winning idea and stating the argument then stating the counter argument to the solution and why it is in fact inferior. A sum up of the pros and cons if you please. This is all explained in the cookbook and an example if given of "Chinook sample database." See that example here https://cookbook.openai.com/examples/how_to_call_functions_with_chat_modelsi . Also, Go here https://platform.openai.com/docs/actions/introduction The work must result in a GPT. So all the configuration will be texted to me. I will upload the configuration, into the text boxes of the GPT configurations and confirm that it works in the GPTs store and then pay and likely a tip. NOTE I do not want an api, I am not needing the function calls to go outside the GPT.
The OPENAI website says to officially make a GPT you must do this: To build a GPT with an action, it is important to understand the end-to-end flow.
Create a GPT in the ChatGPT UI(DO NOT CREATE A UI, OPEN AI provides the GUI)
Manually configure or use the GPT builder to create a GPT
Identify the API(s) you want to use
Go to the "Configure" tab in the GPT editor and select "Create new action"
You will be presented with 3 main options: selecting the authentication schema for the action, inputting the schema itself, and setting the privacy policy URL
The Schema follows the OpenAI API specification format (not to be confused with requiring a Api Key, that is only needed in hosting an Assistant for others to use on your own website and your charged per usage)
Fill in the details for the schema, authentication.
GPTs ask the user for their state problem. With that problem it goes into a series of steps.
Step 1: Generate Initial Answers
Description: Generate three potential answers to the user's stated problem.
Implementation: Use the GPT model to propose three distinct solutions based on the initial problem statement.
Step 2: Evaluate and Rank Answers
Description: Evaluate each of the three answers based on five valid criteria such as cost, ingenuity, feasibility, etc., and rank them.
Implementation: Apply the GPT model to assess each solution against the criteria and provide a ranking based on the overall score.
Step 3: Explore Top Answer
Description: Focus on the highest-scoring answer from Step 2. Explore this concept in detail and provide a summary of its pros and cons.
Implementation: Use the GPT model to delve deeper into the best solution, highlighting its advantages and potential drawbacks.
Step 4: Generate Competing Answers and Explore
Description: Generate two new competing answers. Explore all three current solutions (the winner from Step 3 and the two new ideas) in detail and definition.
Implementation: The GPT model generates two additional solutions and then provides a detailed analysis and description of all three current options.
Step 5: Second Evaluation and Ranking
Description: Repeat the grading and ranking process, similar to Step 2, for the three current solutions.
Implementation: The GPT model re-evaluates all three solutions based on the same criteria and ranks them.
Loop Logic: If necessary, the process from Step 4 and Step 5 can be repeated, iterating through the generation of new solutions and re-evaluation to refine the ideas further.
Final Step: Executive Summary and Counterargument
Description: Provide an executive summary of the winning idea from the final iteration and also articulate a cutting criticism or a counterargument against it.
Implementation: The GPT model synthesizes an executive summary of the top solution and develops a cogent counterargument to present an alternative perspective.
This revised TOT strategy ensures a thorough and dynamic approach to problem-solving, where ideas are not only generated and evaluated but also refined and critically analyzed through iterative loops. The final step ensures that the selected idea is both well-understood and its potential weaknesses are considered.
Key Responsibilities:
- Use GPT to perform text generation and data processing
- Ensure machine learning principles are effectively integrated
- Validate that all computations are handled by the GPT and the Actions feature
Ideal Capabilities:
- Strong expertise with Python and GPT (Generative Pretrained Transformer)
- Extensive experience with data processing, machine learning, and text generation
- Proficiency in developing and working with schemas
Tone of the Generated Text:
The text generated by the GPT should bear a professional and formal tone.
Data Handling Instructions:
The project requires specific formatting for certain data. Please ensure these formatting requirements are met during the text generation process.
This task requires a developer with a keen eye for refinement and precision. Your expertise will play a crucial role in the success of the project's goal to shift computation responsibility from just Python code to the GPT and its Actions feature. FINAL NOTE: "Warning: Existing API integrations that use your secret key and do not explicitly specify an OpenAI-Organization header will be affected. Any such API usage will count as usage for the DocMatthew's organization."
Suositellut taidot
Kilpailun parhaat työt
-
achyuttiwari22 India
-
wintersoldier538 India
-
mailtoafaqCEO Pakistan
-
ALIF80 Bangladesh
-
ALIF80 Bangladesh
Julkinen selvennystaulu
Kuinka päästä alkuun kilpailuiden kanssa
-
Ilmoita kilpailusi Nopeaa ja helppoa
-
Vastaanota tonnikaupalla osallistumisia Ympäri maailmaa
-
Myönnä palkinto parhaalle työlle Lataa tiedostot - Helppoa!