yuchenlin / swiftsage Goto Github PK

View Code? Open in Web Editor NEW

225.0 9.0 21.0 7.4 MB

SwiftSage: A Generative Agent with Fast and Slow Thinking for Complex Interactive Tasks

Home Page: https://yuchenlin.xyz/swiftsage/

Python 97.39% Shell 2.61%

agent decision-making gpt4 llm

swiftsage's Issues

It's said that "copy traj_data_1shot from slow_agent to this directory," in baselines/ReAct_baseline/README.md. However, I didn't find that in "slow_agent" folder. How should I generate relative files?

No such file or directory: 'data_v4/fast_system.train.jsonl'

Error when

(swiftsage) $ unzip goldpaths-all.zip 
Archive:  goldpaths-all.zip
  inflating: goldsequences-0-1-2-3-4-5-6-7-8-9-10-11-12-13-14-15-16-17-18-19-20-21-22-23-24-25-26-27-28-29.json  
(swiftsage) $ python data_convert.py
...
...
...
1-4: 1275
7-3: 734
6-1: 721
6-3: 716
7-1: 609
7-2: 609
6-2: 474
8-1: 424
3-2: 344
3-1: 194
8-2: 89
go: 16680
look: 15259
open: 15098
use: 14056
examine: 13597
wait1: 13036
move: 10349
focus: 6432
pick: 5435
connect: 4360
pour: 4046
activate: 2866
deactivate: 2233
0: 1729
drop: 743
wait: 302
mix: 140
read: 32
Traceback (most recent call last):
  File "data_convert.py", line 197, in <module>
    with open(f"{args.dir}/{mode}.train.jsonl", 'w') as f:
FileNotFoundError: [Errno 2] No such file or directory: 'data_v4/fast_system.train.jsonl'

Is it a known issue? Where can i find data_v4?

where is "slow_agent" directory?

Traceback (most recent call last):
File "eval_agent_fast_slow.py", line 13, in
from eval_utils import load_model, findValidActionNew, load_variation, get_model_output, findValidActionWithSystem2, getFilteredValidActions, sbert_search, clean_look, is_action_failed
File "/home/liuxiaoqian/SwiftSage-main/eval_utils.py", line 12, in
from slow_agent.utils import completion_with_backoff, generate_followup_prompt, load_triplets, sample_few_shot
ModuleNotFoundError: No module named 'slow_agent'

Where is "slow_agent/demos.json'?

In the run_eval_fast_slow. sh， it seems to require a directory named as "slow_agent" that contains a .json file. But it is missing.

The result of ReAct in paper cannot be reproduced

Thanks for your great work! I have used your code to reproduce the react's results. However, I found that my results are much lower than the results of your paper. For example, for tasks 8-2, the score in your paper is 8.0. However, I ran three times, and the score is still only 1.3333. This is the running code.

for task in 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
do
    python eval_agent_react.py \
        --task_nums $task \
        --set test_mini \
        --no_stop \
        --env_step_limit 100 \
        --simplification_str easy \
        --prompt_file baselines/ReAct_baseline/prompt.jsonl \
        --output_path ReAct_logs/gpt-4-1106-preview \
        --model_name gpt-4-1106-preview
done

Do you know why it happened?

AttributeError: 'DeepSpeedCPUAdam' object has no attribute 'ds_opt_adam'

I am getting the following error while running ds_train.sh. Any resolution for this?

AttributeError: 'DeepSpeedCPUAdam' object has no attribute 'ds_opt_adam'

Code Error on running evaluation

Following mentioned evaluation in readme, I encountered following error. Let me know if there is a bug in your provided code.

$ CUDA_VISIBLE_DEVICES=0 python eval_agent_fast_slow.py     --task_nums "28"     --set "test_mini"     --seed 42     --debug_var "450"     --gpt_version "gpt-4"     --output_path "fast_slow_logs/tmp_gpt4/"
{'jar_path': None, 'task_nums': '28', 'env_step_limit': 300, 'lm_path': 'yuchenlin/swift_sw', 'simplification_str': 'easy', 'beams': 5, 'max_episode_per_file': 9999, 'mode': 'fast_system', 'set': 'test_mini', 'output_path': 'fast_slow_logs/tmp_gpt4/', 'compose_mode': 'v4', 'model_parallelism_size': 1, 'seed': 42, 'max_input_len': 1024, 'cut_off': True, 'sbert': True, 'no_stop': True, 'slow_agent': True, 'gpt_version': 'gpt-4', 'local_llm': 'none', 'demo_file': 'data_utils/demos.json', 'debug_var': 450}
fast_slow_logs/tmp_gpt4/task28.log
Downloading (…)okenizer_config.json: 100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 2.54k/2.54k [00:00<00:00, 429kB/s]
Downloading spiece.model: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 792k/792k [00:01<00:00, 634kB/s]
Downloading (…)/main/tokenizer.json: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 2.42M/2.42M [00:01<00:00, 1.58MB/s]
Downloading (…)cial_tokens_map.json: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 2.20k/2.20k [00:00<00:00, 1.45MB/s]
Downloading (…)lve/main/config.json: 100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 788/788 [00:00<00:00, 152kB/s]
Downloading pytorch_model.bin: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1.57G/1.57G [02:17<00:00, 11.4MB/s]
Downloading (…)001fa/.gitattributes: 100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 690/690 [00:00<00:00, 178kB/s]
Downloading (…)_Pooling/config.json: 100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 190/190 [00:00<00:00, 120kB/s]
Downloading (…)3bbb8001fa/README.md: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 3.69k/3.69k [00:00<00:00, 2.58MB/s]
Downloading (…)bb8001fa/config.json: 100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 629/629 [00:00<00:00, 444kB/s]
Downloading (…)ce_transformers.json: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 122/122 [00:00<00:00, 93.6kB/s]
Downloading pytorch_model.bin: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 90.9M/90.9M [00:08<00:00, 11.3MB/s]
Downloading (…)nce_bert_config.json: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 53.0/53.0 [00:00<00:00, 10.3kB/s]
Downloading (…)cial_tokens_map.json: 100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 112/112 [00:00<00:00, 110kB/s]
Downloading (…)001fa/tokenizer.json: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 466k/466k [00:00<00:00, 743kB/s]
Downloading (…)okenizer_config.json: 100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 314/314 [00:00<00:00, 301kB/s]
Downloading (…)3bbb8001fa/vocab.txt: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 232k/232k [00:00<00:00, 606kB/s]
Downloading (…)b8001fa/modules.json: 100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 229/229 [00:00<00:00, 227kB/s]
Traceback (most recent call last):
  File "eval_agent_fast_slow.py", line 627, in <module>
    main()
  File "eval_agent_fast_slow.py", line 624, in main
    eval(args, int(task_num), logger)
  File "eval_agent_fast_slow.py", line 402, in eval
    used_sys2, return_result = findValidActionWithSystem2(predStrs, env, task_num, task_description, info['look'],
  File "/SwiftSage/eval_utils.py", line 589, in findValidActionWithSystem2
    messages=[{"role": "user", "content": prompt_to_next_actions},
UnboundLocalError: local variable 'prompt_to_next_actions' referenced before assignment

Recommend Projects

React

A declarative, efficient, and flexible JavaScript library for building user interfaces.
Vue.js

🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
Typescript

TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
TensorFlow

An Open Source Machine Learning Framework for Everyone
Django

The Web framework for perfectionists with deadlines.
Laravel

A PHP framework for web artisans
D3

Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

javascript

JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
web

Some thing interesting about web. New door for the world.
server

A server is a program made to process requests and deliver data to clients.
Machine learning

Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Visualization

Some thing interesting about visualization, use data art
Game

Some thing interesting about game, make everyone happy.

Recommend Org

Facebook

We are working to build community through open source technology. NB: members must have two-factor auth.
Microsoft

Open source projects and samples from Microsoft.
Google

Google ❤️ Open Source for everyone.
Alibaba

Alibaba Open Source for everyone
D3

Data-Driven Documents codes.
Tencent

China tencent open source team.