Hugging Face's logo

Datasets:
SunSec
/
ds_jiuzhang

Modalities:
Image
Dataset card Data Studio Files
xet
 Community
SunSec commited on
Commit
011b9f4
·
verified ·
1 Parent(s): 8b4ba88

Add files using upload-large-folder tool

Browse files
Files changed (50) hide show
  1. deep_search/sft/3-28_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829_doc_by_itself_QwQ-32B.log +42 -0
  2. deep_search/sft/3-28_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829_doc_by_itself_QwQ-32B_1.log +35 -0
  3. deep_search/sft/3-28_merged_1174_zh_296_sft_1469_doc_by_itself_QwQ-32B.log +109 -0
  4. deep_search/sft/4-1_1.1k_cleaned_data_1097_doc_by_itself_QwQ-32B.log +40 -0
  5. deep_search/sft/4-1_1.1k_cleaned_data_1097_doc_by_itself_QwQ-32B_1.log +81 -0
  6. deep_search/sft/4-1_no_error_data_871_doc_by_itself_QwQ-32B.log +78 -0
  7. deep_search/sft/4-24__doc_by_itself__add_math871.log +4 -0
  8. deep_search/sft/4-24_math_qwq_4524_selected_add_prompt_871_doc_by_itself_qwen7b_sft_871_checkpoint-78_add_math871_after_search.log +44 -0
  9. deep_search/sft/4-24_math_qwq_4524_selected_add_prompt_871_doc_by_itself_qwen7b_sft_871_checkpoint-78_add_math871_after_search_1.log +36 -0
  10. deep_search/sft/4-24_no_error_data_871_doc_by_itself_DeepSeek-R1-Distill-Qwen-7B.log +198 -0
  11. deep_search/sft/4-24_no_error_data_871_doc_by_itself_Qwen2.5-7B-Instruct_add_math871_1.log +27 -0
  12. deep_search/sft/4-24_no_error_data_871_doc_by_itself_Qwen2.5-7B-Instruct_add_math871_2.log +97 -0
  13. deep_search/sft/4-25_no_error_data_871_doc_by_itself_QwQ-32B_1.log +746 -0
  14. deep_search/sft/4-4_no_error_data_871_doc_by_itself_DeepSeek-R1-Distill-Qwen-32.log +115 -0
  15. deep_search/sft/5-5_ablation_ques_domain_filtered_data_738_Qwen2.5-7B-Instruct.log +128 -0
  16. deep_search/sft/5-5_ablation_ques_keywords_filtered_data_727_Qwen2.5-7B-Instruct.log +128 -0
  17. deep_search/sft/5-5_ablation_ques_yiwenci_filtered_data_811_Qwen2.5-7B-Instruct.log +173 -0
  18. deep_search/sft/5-5_ablation_resp_format_1064_random_sample_871_Qwen2.5-7B-Instruct.log +72 -0
  19. deep_search/sft/ds_zero3_offload.json +39 -0
  20. deep_search/sft/gen_data_2.log +4 -0
  21. deep_search/sft/hostfile +2 -0
  22. deep_search/sft/mix.sh +265 -0
  23. deep_search/sft/mix_2.sh +295 -0
  24. deep_search/sft/mix_2_1.sh +174 -0
  25. deep_search/sft/mix_math.sh +62 -0
  26. deep_search/sft/mix_math_after_search.sh +61 -0
  27. deep_search/sft/mix_math_multi_node_1.sh +66 -0
  28. deep_search/sft/mix_math_sht_new_prompt.sh +62 -0
  29. deep_search/sft/mix_wo_mask.sh +62 -0
  30. deep_search/sft/run.sh +25 -0
  31. deep_search/sft/sft.py +204 -0
  32. deep_search/sft/sft_2_math_after_search.py +260 -0
  33. deep_search/sft/sft_2_wo_mask.py +261 -0
  34. deep_search/sft/test.ipynb +58 -0
  35. deep_search/sft/test_len.sh +25 -0
  36. deep_search/sft/test_two_model_qwq.sh +64 -0
  37. deep_search/sft/test_two_model_qwq_1.sh +69 -0
  38. deep_search/sft/train-inst.py +311 -0
  39. deep_search/sft/train-inst.sh +56 -0
  40. deep_search/sft/train_env.yml +145 -0
  41. deep_search/sft/train_requirements.txt +116 -0
  42. deep_search/sft/wait_eval_1.py +167 -0
  43. deep_search/sft/wait_eval_use_one_model_for_ckpt.py +178 -0
  44. deep_search/sft/wait_eval_use_one_model_for_ckpt_two_model.py +178 -0
  45. deep_search/sft/wait_eval_use_one_model_for_ckpt_two_model_dpsk.py +178 -0
  46. deep_search/sft/wait_eval_use_one_model_for_ckpt_two_model_qwq.py +178 -0
  47. deep_search/sft/wait_eval_use_one_model_for_ckpt_worker0.py +180 -0
  48. ssh_node.sh +5 -0
  49. test.py +58 -0
  50. train_requirements.txt +122 -0
deep_search/sft/3-28_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829_doc_by_itself_QwQ-32B.log ADDED
@@ -0,0 +1,42 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ nohup: ignoring input
2
+ step_list: [168, 143, 115, 86]
3
+ step_list: [168, 143, 115, 86]
4
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-168) to exist...
5
+ The checkpoint exists. Waiting for running...
6
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/168/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-168 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/168/eval/inf.log 2>&1 &
7
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/168/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-168 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/168/realqa/inf.log 2>&1 &
8
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-143) to exist...
9
+ The checkpoint exists. Waiting for running...
10
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/143/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-143 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/143/eval/inf.log 2>&1 &
11
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/143/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-143 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/143/realqa/inf.log 2>&1 &
12
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-115) to exist...
13
+ The checkpoint exists. Waiting for running...
14
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/115/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-115 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/115/eval/inf.log 2>&1 &
15
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/115/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-115 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/115/realqa/inf.log 2>&1 &
16
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-86) to exist...
17
+ The checkpoint exists. Waiting for running...
18
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/86/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-86 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/86/eval/inf.log 2>&1 &
19
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/86/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-86 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/86/realqa/inf.log 2>&1 &
20
+ All checkpoints exist. Wait for runing...
21
+ available_gpus: [0, 1, 2, 3, 4, 5, 6, 7]
22
+ The following command is about to run:
23
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=0,1 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/168/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-168 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/168/eval/inf.log 2>&1 &
24
+ The following command is about to run:
25
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,3 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/168/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-168 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/168/realqa/inf.log 2>&1 &
26
+ The following command is about to run:
27
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/143/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-143 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/143/eval/inf.log 2>&1 &
28
+ The following command is about to run:
29
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/143/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-143 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/143/realqa/inf.log 2>&1 &
30
+ available_gpus: [4, 5]
31
+ The following command is about to run:
32
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/115/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-115 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/115/eval/inf.log 2>&1 &
33
+ available_gpus: [4, 5]
34
+ The following command is about to run:
35
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/115/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-115 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/115/realqa/inf.log 2>&1 &
36
+ available_gpus: [4, 5]
37
+ The following command is about to run:
38
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/86/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-86 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/86/eval/inf.log 2>&1 &
39
+ available_gpus: [4, 5]
40
+ The following command is about to run:
41
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/86/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-86 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/86/realqa/inf.log 2>&1 &
42
+ Wish me good luck!
deep_search/sft/3-28_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829_doc_by_itself_QwQ-32B_1.log ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ nohup: ignoring input
2
+ step_list: [168, 143, 115, 86]
3
+ step_list: [168, 143, 115, 86]
4
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-168) to exist...
5
+ The checkpoint exists. Waiting for running...
6
+ skip evaluated model: JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/168/eval
7
+ skip evaluated model: JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/168/realqa
8
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-143) to exist...
9
+ The checkpoint exists. Waiting for running...
10
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/143/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-143 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/143/eval/inf.log 2>&1 &
11
+ skip evaluated model: JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/143/realqa
12
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-115) to exist...
13
+ The checkpoint exists. Waiting for running...
14
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/115/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-115 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/115/eval/inf.log 2>&1 &
15
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/115/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-115 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/115/realqa/inf.log 2>&1 &
16
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-86) to exist...
17
+ The checkpoint exists. Waiting for running...
18
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/86/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-86 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/86/eval/inf.log 2>&1 &
19
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/86/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-86 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/86/realqa/inf.log 2>&1 &
20
+ All checkpoints exist. Wait for runing...
21
+ available_gpus: [2, 3, 4, 6, 7]
22
+ The following command is about to run:
23
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,3 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/143/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-143 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/143/eval/inf.log 2>&1 &
24
+ The following command is about to run:
25
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,6 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/115/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-115 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/115/eval/inf.log 2>&1 &
26
+ available_gpus: [4, 6, 7]
27
+ The following command is about to run:
28
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,6 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/115/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-115 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/115/realqa/inf.log 2>&1 &
29
+ available_gpus: [4, 6, 7]
30
+ The following command is about to run:
31
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,6 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/86/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-86 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/86/eval/inf.log 2>&1 &
32
+ available_gpus: [4, 6, 7]
33
+ The following command is about to run:
34
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,6 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/86/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/checkpoint-86 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829/86/realqa/inf.log 2>&1 &
35
+ Wish me good luck!
deep_search/sft/3-28_merged_1174_zh_296_sft_1469_doc_by_itself_QwQ-32B.log ADDED
@@ -0,0 +1,109 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ nohup: ignoring input
2
+ step_list: [138, 115, 92, 69]
3
+ step_list: [138, 115, 92, 69]
4
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/checkpoint-138) to exist...
5
+ The checkpoint exists. Waiting for running...
6
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=0,2 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/138/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/checkpoint-138 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/138/eval/inf.log 2>&1 &
7
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/138/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/checkpoint-138 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/138/realqa/inf.log 2>&1 &
8
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/checkpoint-115) to exist...
9
+ The checkpoint exists. Waiting for running...
10
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/115/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/checkpoint-115 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/115/eval/inf.log 2>&1 &
11
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/115/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/checkpoint-115 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/115/realqa/inf.log 2>&1 &
12
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/checkpoint-92) to exist...
13
+ The checkpoint exists. Waiting for running...
14
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/92/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/checkpoint-92 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/92/eval/inf.log 2>&1 &
15
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=0,4 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/92/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/checkpoint-92 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/92/realqa/inf.log 2>&1 &
16
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/checkpoint-69) to exist...
17
+ The checkpoint exists. Waiting for running...
18
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/69/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/checkpoint-69 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/69/eval/inf.log 2>&1 &
19
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/69/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/checkpoint-69 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/69/realqa/inf.log 2>&1 &
20
+ All checkpoints exist. Wait for runing...
21
+ available_gpus: [0, 1, 2, 3, 4, 5, 6, 7]
22
+ The following command is about to run:
23
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=0,1 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/138/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/checkpoint-138 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/138/eval/inf.log 2>&1 &
24
+ The following command is about to run:
25
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,3 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/138/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/checkpoint-138 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/138/realqa/inf.log 2>&1 &
26
+ The following command is about to run:
27
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/115/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/checkpoint-115 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/115/eval/inf.log 2>&1 &
28
+ The following command is about to run:
29
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/115/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/checkpoint-115 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/115/realqa/inf.log 2>&1 &
30
+ available_gpus: []
31
+ available_gpus: [1]
32
+ available_gpus: [1]
33
+ available_gpus: [1]
34
+ available_gpus: [1]
35
+ available_gpus: [1]
36
+ available_gpus: [1]
37
+ available_gpus: [1]
38
+ available_gpus: [1]
39
+ available_gpus: [1]
40
+ available_gpus: [1]
41
+ available_gpus: [1]
42
+ available_gpus: [1]
43
+ available_gpus: [1]
44
+ available_gpus: [1]
45
+ available_gpus: [1]
46
+ available_gpus: [1]
47
+ available_gpus: [1]
48
+ available_gpus: [1]
49
+ available_gpus: [1]
50
+ available_gpus: [1]
51
+ available_gpus: [1]
52
+ available_gpus: [1]
53
+ available_gpus: [1]
54
+ available_gpus: [1]
55
+ available_gpus: [1]
56
+ available_gpus: [1]
57
+ available_gpus: [1]
58
+ available_gpus: [1]
59
+ available_gpus: [1]
60
+ available_gpus: [1]
61
+ available_gpus: [1]
62
+ available_gpus: [1]
63
+ available_gpus: [1]
64
+ available_gpus: [1]
65
+ available_gpus: [1]
66
+ available_gpus: [1]
67
+ available_gpus: [1]
68
+ available_gpus: [1]
69
+ available_gpus: [1]
70
+ available_gpus: [1]
71
+ available_gpus: [1]
72
+ available_gpus: [1]
73
+ available_gpus: [1, 4, 5]
74
+ The following command is about to run:
75
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=1,4 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/92/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/checkpoint-92 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/92/eval/inf.log 2>&1 &
76
+ available_gpus: [1, 4, 5]
77
+ The following command is about to run:
78
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=1,4 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/92/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/checkpoint-92 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/92/realqa/inf.log 2>&1 &
79
+ available_gpus: [5]
80
+ available_gpus: [5]
81
+ available_gpus: [5]
82
+ available_gpus: [5]
83
+ available_gpus: [5]
84
+ available_gpus: [5]
85
+ available_gpus: [5]
86
+ available_gpus: [5]
87
+ available_gpus: [5]
88
+ available_gpus: [5]
89
+ available_gpus: [5]
90
+ available_gpus: [5]
91
+ available_gpus: [5]
92
+ available_gpus: [5]
93
+ available_gpus: [5]
94
+ available_gpus: [5]
95
+ available_gpus: [5]
96
+ available_gpus: [5]
97
+ available_gpus: [5]
98
+ available_gpus: [5]
99
+ available_gpus: [5]
100
+ available_gpus: [5]
101
+ available_gpus: [5]
102
+ available_gpus: [0, 2, 5]
103
+ The following command is about to run:
104
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=0,2 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/69/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/checkpoint-69 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/69/eval/inf.log 2>&1 &
105
+ available_gpus: [5]
106
+ available_gpus: [5, 6, 7]
107
+ The following command is about to run:
108
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=5,6 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/69/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/checkpoint-69 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469/69/realqa/inf.log 2>&1 &
109
+ Wish me good luck!
deep_search/sft/4-1_1.1k_cleaned_data_1097_doc_by_itself_QwQ-32B.log ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ step_list: [102, 86, 69, 51]
2
+ step_list: [102, 86, 69, 51]
3
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-102) to exist...
4
+ The checkpoint exists. Waiting for running...
5
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/102/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-102 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/102/eval/inf.log 2>&1 &
6
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/102/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-102 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/102/realqa/inf.log 2>&1 &
7
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-86) to exist...
8
+ The checkpoint exists. Waiting for running...
9
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/86/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-86 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/86/eval/inf.log 2>&1 &
10
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/86/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-86 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/86/realqa/inf.log 2>&1 &
11
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-69) to exist...
12
+ The checkpoint exists. Waiting for running...
13
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/69/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-69 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/69/eval/inf.log 2>&1 &
14
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/69/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-69 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/69/realqa/inf.log 2>&1 &
15
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-51) to exist...
16
+ The checkpoint exists. Waiting for running...
17
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/51/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-51 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/51/eval/inf.log 2>&1 &
18
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/51/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-51 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/51/realqa/inf.log 2>&1 &
19
+ All checkpoints exist. Wait for runing...
20
+ available_gpus: [0, 1, 2, 4, 5, 7]
21
+ The following command is about to run:
22
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=0,1 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/102/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-102 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/102/eval/inf.log 2>&1 &
23
+ The following command is about to run:
24
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,4 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/102/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-102 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/102/realqa/inf.log 2>&1 &
25
+ The following command is about to run:
26
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=5,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/86/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-86 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/86/eval/inf.log 2>&1 &
27
+ available_gpus: [0, 1, 2, 4, 5, 7]
28
+ The following command is about to run:
29
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=0,1 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/86/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-86 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/86/realqa/inf.log 2>&1 &
30
+ The following command is about to run:
31
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,4 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/69/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-69 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/69/eval/inf.log 2>&1 &
32
+ The following command is about to run:
33
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=5,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/69/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-69 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/69/realqa/inf.log 2>&1 &
34
+ available_gpus: []
35
+ available_gpus: [0, 1, 2, 4]
36
+ The following command is about to run:
37
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=0,1 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/51/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-51 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/51/eval/inf.log 2>&1 &
38
+ The following command is about to run:
39
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,4 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/51/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-51 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/51/realqa/inf.log 2>&1 &
40
+ Wish me good luck!
deep_search/sft/4-1_1.1k_cleaned_data_1097_doc_by_itself_QwQ-32B_1.log ADDED
@@ -0,0 +1,81 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ nohup: ignoring input
2
+ step_list: [102, 86, 69, 51]
3
+ step_list: [102, 86, 69, 51]
4
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-102) to exist...
5
+ The checkpoint exists. Waiting for running...
6
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/102/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-102 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/102/eval/inf.log 2>&1 &
7
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/102/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-102 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/102/realqa/inf.log 2>&1 &
8
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-86) to exist...
9
+ The checkpoint exists. Waiting for running...
10
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/86/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-86 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/86/eval/inf.log 2>&1 &
11
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/86/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-86 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/86/realqa/inf.log 2>&1 &
12
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-69) to exist...
13
+ The checkpoint exists. Waiting for running...
14
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/69/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-69 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/69/eval/inf.log 2>&1 &
15
+ skip evaluated model: JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/69/realqa
16
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-51) to exist...
17
+ The checkpoint exists. Waiting for running...
18
+ skip evaluated model: JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/51/eval
19
+ skip evaluated model: JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/51/realqa
20
+ All checkpoints exist. Wait for runing...
21
+ available_gpus: [0, 1, 2, 4, 5, 7]
22
+ The following command is about to run:
23
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=0,1 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/102/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-102 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/102/eval/inf.log 2>&1 &
24
+ The following command is about to run:
25
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,4 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/102/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-102 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/102/realqa/inf.log 2>&1 &
26
+ The following command is about to run:
27
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=5,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/86/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-86 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/86/eval/inf.log 2>&1 &
28
+ available_gpus: []
29
+ available_gpus: []
30
+ available_gpus: []
31
+ available_gpus: []
32
+ available_gpus: []
33
+ available_gpus: []
34
+ available_gpus: []
35
+ available_gpus: []
36
+ available_gpus: []
37
+ available_gpus: []
38
+ available_gpus: []
39
+ available_gpus: []
40
+ available_gpus: []
41
+ available_gpus: []
42
+ available_gpus: []
43
+ available_gpus: []
44
+ available_gpus: []
45
+ available_gpus: []
46
+ available_gpus: []
47
+ available_gpus: []
48
+ available_gpus: []
49
+ available_gpus: []
50
+ available_gpus: []
51
+ available_gpus: []
52
+ available_gpus: []
53
+ available_gpus: []
54
+ available_gpus: []
55
+ available_gpus: []
56
+ available_gpus: []
57
+ available_gpus: []
58
+ available_gpus: []
59
+ available_gpus: []
60
+ available_gpus: []
61
+ available_gpus: []
62
+ available_gpus: []
63
+ available_gpus: []
64
+ available_gpus: []
65
+ available_gpus: []
66
+ available_gpus: []
67
+ available_gpus: []
68
+ available_gpus: []
69
+ available_gpus: []
70
+ available_gpus: []
71
+ available_gpus: []
72
+ available_gpus: []
73
+ available_gpus: [2, 4]
74
+ The following command is about to run:
75
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,4 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/86/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-86 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/86/realqa/inf.log 2>&1 &
76
+ available_gpus: []
77
+ available_gpus: []
78
+ available_gpus: [0, 1]
79
+ The following command is about to run:
80
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=0,1 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/69/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/checkpoint-69 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097/69/eval/inf.log 2>&1 &
81
+ Wish me good luck!
deep_search/sft/4-1_no_error_data_871_doc_by_itself_QwQ-32B.log ADDED
@@ -0,0 +1,78 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ step_list: [78, 68, 55, 41]
2
+ step_list: [78, 68, 55, 41]
3
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/checkpoint-78) to exist...
4
+ The checkpoint exists. Waiting for running...
5
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/78/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/checkpoint-78 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/78/eval/inf.log 2>&1 &
6
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/78/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/checkpoint-78 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/78/realqa/inf.log 2>&1 &
7
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/checkpoint-68) to exist...
8
+ The checkpoint exists. Waiting for running...
9
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/68/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/checkpoint-68 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/68/eval/inf.log 2>&1 &
10
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/68/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/checkpoint-68 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/68/realqa/inf.log 2>&1 &
11
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/checkpoint-55) to exist...
12
+ The checkpoint exists. Waiting for running...
13
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/55/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/checkpoint-55 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/55/eval/inf.log 2>&1 &
14
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/55/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/checkpoint-55 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/55/realqa/inf.log 2>&1 &
15
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/checkpoint-41) to exist...
16
+ The checkpoint exists. Waiting for running...
17
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/41/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/checkpoint-41 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/41/eval/inf.log 2>&1 &
18
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/41/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/checkpoint-41 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/41/realqa/inf.log 2>&1 &
19
+ All checkpoints exist. Wait for runing...
20
+ available_gpus: [0, 1, 2, 4, 5, 7]
21
+ The following command is about to run:
22
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=0,1 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/78/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/checkpoint-78 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/78/eval/inf.log 2>&1 &
23
+ The following command is about to run:
24
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,4 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/78/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/checkpoint-78 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/78/realqa/inf.log 2>&1 &
25
+ The following command is about to run:
26
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=5,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/68/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/checkpoint-68 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/68/eval/inf.log 2>&1 &
27
+ available_gpus: [0, 1, 2, 4, 5, 7]
28
+ The following command is about to run:
29
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=0,1 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/68/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/checkpoint-68 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/68/realqa/inf.log 2>&1 &
30
+ The following command is about to run:
31
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,4 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/55/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/checkpoint-55 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/55/eval/inf.log 2>&1 &
32
+ The following command is about to run:
33
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=5,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/55/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/checkpoint-55 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/55/realqa/inf.log 2>&1 &
34
+ available_gpus: []
35
+ available_gpus: []
36
+ available_gpus: []
37
+ available_gpus: []
38
+ available_gpus: []
39
+ available_gpus: []
40
+ available_gpus: []
41
+ available_gpus: []
42
+ available_gpus: []
43
+ available_gpus: []
44
+ available_gpus: []
45
+ available_gpus: []
46
+ available_gpus: []
47
+ available_gpus: []
48
+ available_gpus: []
49
+ available_gpus: []
50
+ available_gpus: []
51
+ available_gpus: []
52
+ available_gpus: []
53
+ available_gpus: []
54
+ available_gpus: []
55
+ available_gpus: []
56
+ available_gpus: []
57
+ available_gpus: []
58
+ available_gpus: []
59
+ available_gpus: []
60
+ available_gpus: []
61
+ available_gpus: []
62
+ available_gpus: []
63
+ available_gpus: []
64
+ available_gpus: []
65
+ available_gpus: []
66
+ available_gpus: []
67
+ available_gpus: []
68
+ available_gpus: [5, 7]
69
+ The following command is about to run:
70
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=5,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/41/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/checkpoint-41 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/41/eval/inf.log 2>&1 &
71
+ available_gpus: []
72
+ available_gpus: []
73
+ available_gpus: []
74
+ available_gpus: [1]
75
+ available_gpus: [0, 1]
76
+ The following command is about to run:
77
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=0,1 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name realqa --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/41/realqa --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/checkpoint-41 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:2284#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_no_error_data_871/41/realqa/inf.log 2>&1 &
78
+ Wish me good luck!
deep_search/sft/4-24__doc_by_itself__add_math871.log ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ Traceback (most recent call last):
2
+ File "/opt/aps/workdir/sunshuang/deep_search/sft/wait_eval_use_one_model_for_ckpt_two_model.py", line 176, in <module>
3
+ ckpt = sys.argv[1]
4
+ IndexError: list index out of range
deep_search/sft/4-24_math_qwq_4524_selected_add_prompt_871_doc_by_itself_qwen7b_sft_871_checkpoint-78_add_math871_after_search.log ADDED
@@ -0,0 +1,44 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ nohup: ignoring input
2
+ step_list: [78, 68, 55, 41]
3
+ step_list: [78, 68, 55, 41]
4
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-78) to exist...
5
+ The checkpoint exists. Waiting for running...
6
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/78/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-78 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/78/eval/inf.log 2>&1 &
7
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name aime --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/78/aime --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-78 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/78/aime/inf.log 2>&1 &
8
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-68) to exist...
9
+ The checkpoint exists. Waiting for running...
10
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/68/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-68 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/68/eval/inf.log 2>&1 &
11
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name aime --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/68/aime --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-68 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/68/aime/inf.log 2>&1 &
12
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-55) to exist...
13
+ The checkpoint exists. Waiting for running...
14
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/55/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-55 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/55/eval/inf.log 2>&1 &
15
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name aime --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/55/aime --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-55 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/55/aime/inf.log 2>&1 &
16
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-41) to exist...
17
+ The checkpoint exists. Waiting for running...
18
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/41/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-41 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/41/eval/inf.log 2>&1 &
19
+ skip evaluated model: JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/41/aime
20
+ All checkpoints exist. Wait for runing...
21
+ available_gpus: [2, 3, 4, 5, 6, 7]
22
+ The following command is about to run:
23
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,3 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/78/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-78 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/78/eval/inf.log 2>&1 &
24
+ The following command is about to run:
25
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name aime --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/78/aime --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-78 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/78/aime/inf.log 2>&1 &
26
+ The following command is about to run:
27
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/68/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-68 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/68/eval/inf.log 2>&1 &
28
+ available_gpus: [2, 3, 4, 5, 6, 7]
29
+ The following command is about to run:
30
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,3 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name aime --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/68/aime --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-68 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/68/aime/inf.log 2>&1 &
31
+ The following command is about to run:
32
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/55/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-55 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/55/eval/inf.log 2>&1 &
33
+ The following command is about to run:
34
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name aime --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/55/aime --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-55 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/55/aime/inf.log 2>&1 &
35
+ available_gpus: []
36
+ available_gpus: []
37
+ available_gpus: []
38
+ available_gpus: []
39
+ available_gpus: []
40
+ available_gpus: []
41
+ available_gpus: [4, 5]
42
+ The following command is about to run:
43
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/41/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-41 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/41/eval/inf.log 2>&1 &
44
+ Wish me good luck!
deep_search/sft/4-24_math_qwq_4524_selected_add_prompt_871_doc_by_itself_qwen7b_sft_871_checkpoint-78_add_math871_after_search_1.log ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ nohup: ignoring input
2
+ step_list: [78, 68, 55, 41]
3
+ step_list: [78, 68, 55, 41]
4
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-78) to exist...
5
+ The checkpoint exists. Waiting for running...
6
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/78/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-78 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/78/gaia/inf.log 2>&1 &
7
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-68) to exist...
8
+ The checkpoint exists. Waiting for running...
9
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/68/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-68 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/68/gaia/inf.log 2>&1 &
10
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-55) to exist...
11
+ The checkpoint exists. Waiting for running...
12
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/55/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-55 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/55/gaia/inf.log 2>&1 &
13
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-41) to exist...
14
+ The checkpoint exists. Waiting for running...
15
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/41/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-41 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/41/gaia/inf.log 2>&1 &
16
+ All checkpoints exist. Wait for runing...
17
+ available_gpus: [2, 3, 4, 5, 6, 7]
18
+ The following command is about to run:
19
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,3 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/78/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-78 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/78/gaia/inf.log 2>&1 &
20
+ The following command is about to run:
21
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/68/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-68 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/68/gaia/inf.log 2>&1 &
22
+ The following command is about to run:
23
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/55/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-55 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/55/gaia/inf.log 2>&1 &
24
+ available_gpus: []
25
+ available_gpus: []
26
+ available_gpus: []
27
+ available_gpus: []
28
+ available_gpus: []
29
+ available_gpus: []
30
+ available_gpus: []
31
+ available_gpus: []
32
+ available_gpus: []
33
+ available_gpus: [4, 5]
34
+ The following command is about to run:
35
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/41/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-41 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/41/gaia/inf.log 2>&1 &
36
+ Wish me good luck!
deep_search/sft/4-24_no_error_data_871_doc_by_itself_DeepSeek-R1-Distill-Qwen-7B.log ADDED
@@ -0,0 +1,198 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ nohup: ignoring input
2
+ step_list: [78, 68, 55, 41]
3
+ step_list: [78, 68, 55, 41]
4
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/checkpoint-78) to exist...
5
+ The checkpoint exists. Waiting for running...
6
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_dpsk.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/78/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/checkpoint-78 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/78/eval/inf.log 2>&1 &
7
+ skip evaluated model: JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/78/gaia
8
+ skip evaluated model: JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/78/aime
9
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/checkpoint-68) to exist...
10
+ The checkpoint exists. Waiting for running...
11
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_dpsk.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/68/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/checkpoint-68 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/68/eval/inf.log 2>&1 &
12
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_dpsk.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/68/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/checkpoint-68 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/68/gaia/inf.log 2>&1 &
13
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_dpsk.py --dataset_name aime --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/68/aime --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/checkpoint-68 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/68/aime/inf.log 2>&1 &
14
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/checkpoint-55) to exist...
15
+ The checkpoint exists. Waiting for running...
16
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_dpsk.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/55/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/checkpoint-55 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/55/eval/inf.log 2>&1 &
17
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_dpsk.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/55/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/checkpoint-55 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/55/gaia/inf.log 2>&1 &
18
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_dpsk.py --dataset_name aime --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/55/aime --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/checkpoint-55 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/55/aime/inf.log 2>&1 &
19
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/checkpoint-41) to exist...
20
+ The checkpoint exists. Waiting for running...
21
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_dpsk.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/41/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/checkpoint-41 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/41/eval/inf.log 2>&1 &
22
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_dpsk.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/41/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/checkpoint-41 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/41/gaia/inf.log 2>&1 &
23
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_dpsk.py --dataset_name aime --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/41/aime --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/checkpoint-41 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/41/aime/inf.log 2>&1 &
24
+ All checkpoints exist. Wait for runing...
25
+ available_gpus: [2, 3, 4, 5, 6, 7]
26
+ The following command is about to run:
27
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,3 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_dpsk.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/78/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/checkpoint-78 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/78/eval/inf.log 2>&1 &
28
+ The following command is about to run:
29
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_dpsk.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/68/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/checkpoint-68 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/68/eval/inf.log 2>&1 &
30
+ The following command is about to run:
31
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_dpsk.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/68/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/checkpoint-68 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/68/gaia/inf.log 2>&1 &
32
+ available_gpus: []
33
+ available_gpus: []
34
+ available_gpus: []
35
+ available_gpus: []
36
+ available_gpus: []
37
+ available_gpus: []
38
+ available_gpus: []
39
+ available_gpus: []
40
+ available_gpus: []
41
+ available_gpus: []
42
+ available_gpus: []
43
+ available_gpus: []
44
+ available_gpus: []
45
+ available_gpus: []
46
+ available_gpus: []
47
+ available_gpus: []
48
+ available_gpus: []
49
+ available_gpus: []
50
+ available_gpus: []
51
+ available_gpus: []
52
+ available_gpus: []
53
+ available_gpus: []
54
+ available_gpus: []
55
+ available_gpus: []
56
+ available_gpus: []
57
+ available_gpus: []
58
+ available_gpus: []
59
+ available_gpus: []
60
+ available_gpus: [6, 7]
61
+ The following command is about to run:
62
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_dpsk.py --dataset_name aime --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/68/aime --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/checkpoint-68 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/68/aime/inf.log 2>&1 &
63
+ available_gpus: []
64
+ available_gpus: []
65
+ available_gpus: []
66
+ available_gpus: []
67
+ available_gpus: []
68
+ available_gpus: []
69
+ available_gpus: []
70
+ available_gpus: []
71
+ available_gpus: [6, 7]
72
+ The following command is about to run:
73
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_dpsk.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/55/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/checkpoint-55 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/55/eval/inf.log 2>&1 &
74
+ available_gpus: []
75
+ available_gpus: []
76
+ available_gpus: []
77
+ available_gpus: []
78
+ available_gpus: []
79
+ available_gpus: []
80
+ available_gpus: []
81
+ available_gpus: []
82
+ available_gpus: []
83
+ available_gpus: []
84
+ available_gpus: []
85
+ available_gpus: []
86
+ available_gpus: []
87
+ available_gpus: []
88
+ available_gpus: []
89
+ available_gpus: []
90
+ available_gpus: []
91
+ available_gpus: []
92
+ available_gpus: []
93
+ available_gpus: []
94
+ available_gpus: []
95
+ available_gpus: []
96
+ available_gpus: []
97
+ available_gpus: []
98
+ available_gpus: []
99
+ available_gpus: []
100
+ available_gpus: []
101
+ available_gpus: []
102
+ available_gpus: []
103
+ available_gpus: []
104
+ available_gpus: []
105
+ available_gpus: []
106
+ available_gpus: []
107
+ available_gpus: []
108
+ available_gpus: []
109
+ available_gpus: []
110
+ available_gpus: []
111
+ available_gpus: []
112
+ available_gpus: []
113
+ available_gpus: []
114
+ available_gpus: []
115
+ available_gpus: []
116
+ available_gpus: []
117
+ available_gpus: []
118
+ available_gpus: []
119
+ available_gpus: []
120
+ available_gpus: []
121
+ available_gpus: []
122
+ available_gpus: []
123
+ available_gpus: []
124
+ available_gpus: []
125
+ available_gpus: []
126
+ available_gpus: []
127
+ available_gpus: []
128
+ available_gpus: []
129
+ available_gpus: []
130
+ available_gpus: []
131
+ available_gpus: []
132
+ available_gpus: []
133
+ available_gpus: []
134
+ available_gpus: []
135
+ available_gpus: []
136
+ available_gpus: []
137
+ available_gpus: []
138
+ available_gpus: []
139
+ available_gpus: []
140
+ available_gpus: []
141
+ available_gpus: []
142
+ available_gpus: []
143
+ available_gpus: []
144
+ available_gpus: []
145
+ available_gpus: []
146
+ available_gpus: []
147
+ available_gpus: []
148
+ available_gpus: [4, 5]
149
+ The following command is about to run:
150
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_dpsk.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/55/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/checkpoint-55 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/55/gaia/inf.log 2>&1 &
151
+ available_gpus: [4, 5]
152
+ The following command is about to run:
153
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_dpsk.py --dataset_name aime --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/55/aime --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/checkpoint-55 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/55/aime/inf.log 2>&1 &
154
+ available_gpus: []
155
+ available_gpus: []
156
+ available_gpus: []
157
+ available_gpus: []
158
+ available_gpus: []
159
+ available_gpus: [2, 3]
160
+ The following command is about to run:
161
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,3 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_dpsk.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/41/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/checkpoint-41 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/41/eval/inf.log 2>&1 &
162
+ available_gpus: []
163
+ available_gpus: []
164
+ available_gpus: []
165
+ available_gpus: []
166
+ available_gpus: []
167
+ available_gpus: []
168
+ available_gpus: []
169
+ available_gpus: []
170
+ available_gpus: []
171
+ available_gpus: []
172
+ available_gpus: []
173
+ available_gpus: []
174
+ available_gpus: []
175
+ available_gpus: []
176
+ available_gpus: []
177
+ available_gpus: []
178
+ available_gpus: []
179
+ available_gpus: []
180
+ available_gpus: []
181
+ available_gpus: []
182
+ available_gpus: []
183
+ available_gpus: []
184
+ available_gpus: []
185
+ available_gpus: []
186
+ available_gpus: []
187
+ available_gpus: []
188
+ available_gpus: []
189
+ available_gpus: []
190
+ available_gpus: []
191
+ available_gpus: []
192
+ available_gpus: [6, 7]
193
+ The following command is about to run:
194
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_dpsk.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/41/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/checkpoint-41 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/41/gaia/inf.log 2>&1 &
195
+ available_gpus: [4, 5]
196
+ The following command is about to run:
197
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_dpsk.py --dataset_name aime --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/41/aime --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/checkpoint-41 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:28315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-7B#TOKEN:DeepSeek-R1-Distill-Qwen-7B#BSZ:2#ACC:4_no_error_data_871/41/aime/inf.log 2>&1 &
198
+ Wish me good luck!
deep_search/sft/4-24_no_error_data_871_doc_by_itself_Qwen2.5-7B-Instruct_add_math871_1.log ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ nohup: ignoring input
2
+ step_list: [162, 136, 109, 81]
3
+ step_list: [162, 136, 109, 81]
4
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/checkpoint-162) to exist...
5
+ The checkpoint exists. Waiting for running...
6
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/162/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/checkpoint-162 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/162/gaia/inf.log 2>&1 &
7
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/checkpoint-136) to exist...
8
+ The checkpoint exists. Waiting for running...
9
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/136/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/checkpoint-136 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/136/gaia/inf.log 2>&1 &
10
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/checkpoint-109) to exist...
11
+ The checkpoint exists. Waiting for running...
12
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/109/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/checkpoint-109 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/109/gaia/inf.log 2>&1 &
13
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/checkpoint-81) to exist...
14
+ The checkpoint exists. Waiting for running...
15
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/81/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/checkpoint-81 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/81/gaia/inf.log 2>&1 &
16
+ All checkpoints exist. Wait for runing...
17
+ available_gpus: [2, 3, 6, 7]
18
+ The following command is about to run:
19
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,3 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/162/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/checkpoint-162 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/162/gaia/inf.log 2>&1 &
20
+ The following command is about to run:
21
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/136/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/checkpoint-136 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/136/gaia/inf.log 2>&1 &
22
+ available_gpus: [2, 3, 6, 7]
23
+ The following command is about to run:
24
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,3 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/109/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/checkpoint-109 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/109/gaia/inf.log 2>&1 &
25
+ The following command is about to run:
26
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/81/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/checkpoint-81 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:11361#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_no_error_data_871/81/gaia/inf.log 2>&1 &
27
+ Wish me good luck!
deep_search/sft/4-24_no_error_data_871_doc_by_itself_Qwen2.5-7B-Instruct_add_math871_2.log ADDED
@@ -0,0 +1,97 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ nohup: ignoring input
2
+ step_list: [78, 68, 55, 41, 27, 13]
3
+ step_list: [78, 68, 55, 41, 27, 13]
4
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-78) to exist...
5
+ The checkpoint exists. Waiting for running...
6
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/78/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-78 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/78/eval/inf.log 2>&1 &
7
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/78/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-78 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/78/gaia/inf.log 2>&1 &
8
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-68) to exist...
9
+ The checkpoint exists. Waiting for running...
10
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/68/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-68 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/68/eval/inf.log 2>&1 &
11
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/68/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-68 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/68/gaia/inf.log 2>&1 &
12
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-55) to exist...
13
+ The checkpoint exists. Waiting for running...
14
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/55/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-55 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/55/eval/inf.log 2>&1 &
15
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/55/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-55 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/55/gaia/inf.log 2>&1 &
16
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-41) to exist...
17
+ The checkpoint exists. Waiting for running...
18
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/41/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-41 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/41/eval/inf.log 2>&1 &
19
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/41/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-41 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/41/gaia/inf.log 2>&1 &
20
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-27) to exist...
21
+ The checkpoint exists. Waiting for running...
22
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/27/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-27 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/27/eval/inf.log 2>&1 &
23
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/27/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-27 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/27/gaia/inf.log 2>&1 &
24
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-13) to exist...
25
+ The checkpoint exists. Waiting for running...
26
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/13/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-13 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/13/eval/inf.log 2>&1 &
27
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/13/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-13 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/13/gaia/inf.log 2>&1 &
28
+ All checkpoints exist. Wait for runing...
29
+ available_gpus: [2, 3, 4, 5, 6, 7]
30
+ The following command is about to run:
31
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,3 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/78/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-78 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/78/eval/inf.log 2>&1 &
32
+ The following command is about to run:
33
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/78/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-78 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/78/gaia/inf.log 2>&1 &
34
+ The following command is about to run:
35
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/68/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-68 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/68/eval/inf.log 2>&1 &
36
+ available_gpus: []
37
+ available_gpus: []
38
+ available_gpus: []
39
+ available_gpus: []
40
+ available_gpus: []
41
+ available_gpus: []
42
+ available_gpus: []
43
+ available_gpus: []
44
+ available_gpus: [4, 5]
45
+ The following command is about to run:
46
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/68/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-68 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/68/gaia/inf.log 2>&1 &
47
+ available_gpus: []
48
+ available_gpus: []
49
+ available_gpus: []
50
+ available_gpus: []
51
+ available_gpus: []
52
+ available_gpus: []
53
+ available_gpus: []
54
+ available_gpus: []
55
+ available_gpus: [4, 5]
56
+ The following command is about to run:
57
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/55/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/checkpoint-55 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871/55/eval/inf.log 2>&1 &
58
+ available_gpus: []
59
+ available_gpus: []
60
+ available_gpus: []
61
+ available_gpus: []
62
+ available_gpus: []
63
+ available_gpus: []
64
+ available_gpus: []
65
+ available_gpus: []
66
+ available_gpus: []
67
+ available_gpus: []
68
+ available_gpus: []
69
+ available_gpus: []
70
+ available_gpus: []
71
+ available_gpus: []
72
+ available_gpus: []
73
+ available_gpus: []
74
+ available_gpus: []
75
+ available_gpus: []
76
+ available_gpus: []
77
+ available_gpus: []
78
+ available_gpus: []
79
+ available_gpus: []
80
+ available_gpus: []
81
+ available_gpus: []
82
+ available_gpus: []
83
+ available_gpus: []
84
+ available_gpus: []
85
+ available_gpus: []
86
+ available_gpus: []
87
+ available_gpus: []
88
+ available_gpus: []
89
+ available_gpus: []
90
+ available_gpus: []
91
+ available_gpus: []
92
+ available_gpus: []
93
+ available_gpus: []
94
+ available_gpus: []
95
+ available_gpus: []
96
+ available_gpus: []
97
+ available_gpus: []
deep_search/sft/4-25_no_error_data_871_doc_by_itself_QwQ-32B_1.log ADDED
@@ -0,0 +1,746 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ nohup: ignoring input
2
+ step_list: [78, 68, 54, 40, 27, 13]
3
+ step_list: [40, 27, 13]
4
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-40) to exist...
5
+ The checkpoint exists. Waiting for running...
6
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/40/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-40 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/40/eval/inf.log 2>&1 &
7
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/40/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-40 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/40/gaia/inf.log 2>&1 &
8
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name aime --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/40/aime --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-40 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/40/aime/inf.log 2>&1 &
9
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name frames --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/40/frames --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-40 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/40/frames/inf.log 2>&1 &
10
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-27) to exist...
11
+ The checkpoint exists. Waiting for running...
12
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/27/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-27 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/27/eval/inf.log 2>&1 &
13
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/27/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-27 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/27/gaia/inf.log 2>&1 &
14
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name aime --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/27/aime --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-27 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/27/aime/inf.log 2>&1 &
15
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name frames --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/27/frames --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-27 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/27/frames/inf.log 2>&1 &
16
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-13) to exist...
17
+ The checkpoint exists. Waiting for running...
18
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/13/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-13 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/13/eval/inf.log 2>&1 &
19
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/13/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-13 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/13/gaia/inf.log 2>&1 &
20
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name aime --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/13/aime --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-13 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/13/aime/inf.log 2>&1 &
21
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name frames --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/13/frames --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-13 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/13/frames/inf.log 2>&1 &
22
+ All checkpoints exist. Wait for runing...
23
+ available_gpus: [2, 3, 4, 5, 6, 7]
24
+ The following command is about to run:
25
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,3 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/40/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-40 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/40/eval/inf.log 2>&1 &
26
+ The following command is about to run:
27
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/40/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-40 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/40/gaia/inf.log 2>&1 &
28
+ The following command is about to run:
29
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name aime --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/40/aime --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-40 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/40/aime/inf.log 2>&1 &
30
+ available_gpus: []
31
+ available_gpus: []
32
+ available_gpus: []
33
+ available_gpus: []
34
+ available_gpus: []
35
+ available_gpus: []
36
+ available_gpus: []
37
+ available_gpus: []
38
+ available_gpus: []
39
+ available_gpus: []
40
+ available_gpus: []
41
+ available_gpus: []
42
+ available_gpus: [6, 7]
43
+ The following command is about to run:
44
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name frames --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/40/frames --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-40 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/40/frames/inf.log 2>&1 &
45
+ available_gpus: []
46
+ available_gpus: []
47
+ available_gpus: []
48
+ available_gpus: []
49
+ available_gpus: []
50
+ available_gpus: []
51
+ available_gpus: []
52
+ available_gpus: []
53
+ available_gpus: []
54
+ available_gpus: []
55
+ available_gpus: []
56
+ available_gpus: []
57
+ available_gpus: []
58
+ available_gpus: []
59
+ available_gpus: []
60
+ available_gpus: []
61
+ available_gpus: []
62
+ available_gpus: []
63
+ available_gpus: []
64
+ available_gpus: []
65
+ available_gpus: []
66
+ available_gpus: []
67
+ available_gpus: []
68
+ available_gpus: []
69
+ available_gpus: []
70
+ available_gpus: []
71
+ available_gpus: []
72
+ available_gpus: []
73
+ available_gpus: []
74
+ available_gpus: []
75
+ available_gpus: []
76
+ available_gpus: []
77
+ available_gpus: []
78
+ available_gpus: []
79
+ available_gpus: []
80
+ available_gpus: []
81
+ available_gpus: []
82
+ available_gpus: []
83
+ available_gpus: []
84
+ available_gpus: []
85
+ available_gpus: []
86
+ available_gpus: []
87
+ available_gpus: []
88
+ available_gpus: []
89
+ available_gpus: []
90
+ available_gpus: []
91
+ available_gpus: []
92
+ available_gpus: []
93
+ available_gpus: []
94
+ available_gpus: []
95
+ available_gpus: []
96
+ available_gpus: []
97
+ available_gpus: []
98
+ available_gpus: []
99
+ available_gpus: []
100
+ available_gpus: []
101
+ available_gpus: []
102
+ available_gpus: []
103
+ available_gpus: []
104
+ available_gpus: []
105
+ available_gpus: []
106
+ available_gpus: []
107
+ available_gpus: []
108
+ available_gpus: []
109
+ available_gpus: []
110
+ available_gpus: []
111
+ available_gpus: []
112
+ available_gpus: []
113
+ available_gpus: []
114
+ available_gpus: []
115
+ available_gpus: []
116
+ available_gpus: []
117
+ available_gpus: []
118
+ available_gpus: []
119
+ available_gpus: []
120
+ available_gpus: []
121
+ available_gpus: []
122
+ available_gpus: []
123
+ available_gpus: []
124
+ available_gpus: []
125
+ available_gpus: []
126
+ available_gpus: []
127
+ available_gpus: []
128
+ available_gpus: []
129
+ available_gpus: []
130
+ available_gpus: []
131
+ available_gpus: []
132
+ available_gpus: []
133
+ available_gpus: []
134
+ available_gpus: []
135
+ available_gpus: []
136
+ available_gpus: []
137
+ available_gpus: []
138
+ available_gpus: []
139
+ available_gpus: [4, 5]
140
+ The following command is about to run:
141
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/27/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-27 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/27/eval/inf.log 2>&1 &
142
+ available_gpus: []
143
+ available_gpus: []
144
+ available_gpus: []
145
+ available_gpus: []
146
+ available_gpus: []
147
+ available_gpus: []
148
+ available_gpus: []
149
+ available_gpus: []
150
+ available_gpus: []
151
+ available_gpus: []
152
+ available_gpus: []
153
+ available_gpus: []
154
+ available_gpus: []
155
+ available_gpus: []
156
+ available_gpus: []
157
+ available_gpus: []
158
+ available_gpus: []
159
+ available_gpus: []
160
+ available_gpus: []
161
+ available_gpus: []
162
+ available_gpus: []
163
+ available_gpus: []
164
+ available_gpus: []
165
+ available_gpus: []
166
+ available_gpus: []
167
+ available_gpus: []
168
+ available_gpus: []
169
+ available_gpus: []
170
+ available_gpus: []
171
+ available_gpus: []
172
+ available_gpus: []
173
+ available_gpus: []
174
+ available_gpus: []
175
+ available_gpus: []
176
+ available_gpus: []
177
+ available_gpus: []
178
+ available_gpus: []
179
+ available_gpus: []
180
+ available_gpus: []
181
+ available_gpus: []
182
+ available_gpus: []
183
+ available_gpus: []
184
+ available_gpus: []
185
+ available_gpus: []
186
+ available_gpus: []
187
+ available_gpus: []
188
+ available_gpus: []
189
+ available_gpus: []
190
+ available_gpus: []
191
+ available_gpus: []
192
+ available_gpus: []
193
+ available_gpus: []
194
+ available_gpus: []
195
+ available_gpus: []
196
+ available_gpus: []
197
+ available_gpus: []
198
+ available_gpus: []
199
+ available_gpus: []
200
+ available_gpus: []
201
+ available_gpus: []
202
+ available_gpus: []
203
+ available_gpus: []
204
+ available_gpus: []
205
+ available_gpus: []
206
+ available_gpus: []
207
+ available_gpus: []
208
+ available_gpus: []
209
+ available_gpus: []
210
+ available_gpus: []
211
+ available_gpus: []
212
+ available_gpus: []
213
+ available_gpus: []
214
+ available_gpus: []
215
+ available_gpus: []
216
+ available_gpus: []
217
+ available_gpus: []
218
+ available_gpus: []
219
+ available_gpus: []
220
+ available_gpus: []
221
+ available_gpus: []
222
+ available_gpus: []
223
+ available_gpus: []
224
+ available_gpus: []
225
+ available_gpus: []
226
+ available_gpus: []
227
+ available_gpus: []
228
+ available_gpus: []
229
+ available_gpus: []
230
+ available_gpus: []
231
+ available_gpus: []
232
+ available_gpus: []
233
+ available_gpus: []
234
+ available_gpus: []
235
+ available_gpus: []
236
+ available_gpus: []
237
+ available_gpus: []
238
+ available_gpus: []
239
+ available_gpus: []
240
+ available_gpus: []
241
+ available_gpus: []
242
+ available_gpus: []
243
+ available_gpus: []
244
+ available_gpus: []
245
+ available_gpus: []
246
+ available_gpus: []
247
+ available_gpus: []
248
+ available_gpus: []
249
+ available_gpus: []
250
+ available_gpus: []
251
+ available_gpus: []
252
+ available_gpus: []
253
+ available_gpus: []
254
+ available_gpus: []
255
+ available_gpus: []
256
+ available_gpus: []
257
+ available_gpus: []
258
+ available_gpus: []
259
+ available_gpus: []
260
+ available_gpus: []
261
+ available_gpus: []
262
+ available_gpus: []
263
+ available_gpus: []
264
+ available_gpus: []
265
+ available_gpus: []
266
+ available_gpus: []
267
+ available_gpus: []
268
+ available_gpus: []
269
+ available_gpus: []
270
+ available_gpus: []
271
+ available_gpus: []
272
+ available_gpus: []
273
+ available_gpus: []
274
+ available_gpus: []
275
+ available_gpus: []
276
+ available_gpus: []
277
+ available_gpus: []
278
+ available_gpus: []
279
+ available_gpus: []
280
+ available_gpus: []
281
+ available_gpus: []
282
+ available_gpus: []
283
+ available_gpus: []
284
+ available_gpus: []
285
+ available_gpus: []
286
+ available_gpus: []
287
+ available_gpus: []
288
+ available_gpus: []
289
+ available_gpus: []
290
+ available_gpus: []
291
+ available_gpus: []
292
+ available_gpus: []
293
+ available_gpus: []
294
+ available_gpus: []
295
+ available_gpus: []
296
+ available_gpus: []
297
+ available_gpus: []
298
+ available_gpus: []
299
+ available_gpus: []
300
+ available_gpus: []
301
+ available_gpus: []
302
+ available_gpus: []
303
+ available_gpus: []
304
+ available_gpus: []
305
+ available_gpus: []
306
+ available_gpus: []
307
+ available_gpus: []
308
+ available_gpus: []
309
+ available_gpus: []
310
+ available_gpus: []
311
+ available_gpus: []
312
+ available_gpus: []
313
+ available_gpus: []
314
+ available_gpus: []
315
+ available_gpus: []
316
+ available_gpus: []
317
+ available_gpus: []
318
+ available_gpus: []
319
+ available_gpus: []
320
+ available_gpus: []
321
+ available_gpus: []
322
+ available_gpus: []
323
+ available_gpus: []
324
+ available_gpus: []
325
+ available_gpus: []
326
+ available_gpus: []
327
+ available_gpus: []
328
+ available_gpus: []
329
+ available_gpus: []
330
+ available_gpus: []
331
+ available_gpus: []
332
+ available_gpus: []
333
+ available_gpus: []
334
+ available_gpus: []
335
+ available_gpus: []
336
+ available_gpus: []
337
+ available_gpus: []
338
+ available_gpus: []
339
+ available_gpus: []
340
+ available_gpus: []
341
+ available_gpus: []
342
+ available_gpus: []
343
+ available_gpus: []
344
+ available_gpus: []
345
+ available_gpus: []
346
+ available_gpus: []
347
+ available_gpus: []
348
+ available_gpus: []
349
+ available_gpus: []
350
+ available_gpus: []
351
+ available_gpus: []
352
+ available_gpus: []
353
+ available_gpus: []
354
+ available_gpus: []
355
+ available_gpus: []
356
+ available_gpus: []
357
+ available_gpus: []
358
+ available_gpus: []
359
+ available_gpus: []
360
+ available_gpus: []
361
+ available_gpus: []
362
+ available_gpus: []
363
+ available_gpus: []
364
+ available_gpus: []
365
+ available_gpus: []
366
+ available_gpus: []
367
+ available_gpus: []
368
+ available_gpus: []
369
+ available_gpus: []
370
+ available_gpus: []
371
+ available_gpus: []
372
+ available_gpus: []
373
+ available_gpus: []
374
+ available_gpus: []
375
+ available_gpus: []
376
+ available_gpus: []
377
+ available_gpus: []
378
+ available_gpus: []
379
+ available_gpus: []
380
+ available_gpus: []
381
+ available_gpus: []
382
+ available_gpus: []
383
+ available_gpus: []
384
+ available_gpus: []
385
+ available_gpus: []
386
+ available_gpus: []
387
+ available_gpus: []
388
+ available_gpus: []
389
+ available_gpus: []
390
+ available_gpus: []
391
+ available_gpus: []
392
+ available_gpus: []
393
+ available_gpus: []
394
+ available_gpus: []
395
+ available_gpus: []
396
+ available_gpus: []
397
+ available_gpus: []
398
+ available_gpus: []
399
+ available_gpus: []
400
+ available_gpus: []
401
+ available_gpus: []
402
+ available_gpus: []
403
+ available_gpus: []
404
+ available_gpus: []
405
+ available_gpus: []
406
+ available_gpus: []
407
+ available_gpus: []
408
+ available_gpus: []
409
+ available_gpus: []
410
+ available_gpus: []
411
+ available_gpus: []
412
+ available_gpus: []
413
+ available_gpus: []
414
+ available_gpus: []
415
+ available_gpus: []
416
+ available_gpus: []
417
+ available_gpus: []
418
+ available_gpus: []
419
+ available_gpus: []
420
+ available_gpus: []
421
+ available_gpus: []
422
+ available_gpus: []
423
+ available_gpus: []
424
+ available_gpus: []
425
+ available_gpus: []
426
+ available_gpus: []
427
+ available_gpus: []
428
+ available_gpus: []
429
+ available_gpus: []
430
+ available_gpus: []
431
+ available_gpus: []
432
+ available_gpus: []
433
+ available_gpus: []
434
+ available_gpus: []
435
+ available_gpus: []
436
+ available_gpus: []
437
+ available_gpus: []
438
+ available_gpus: []
439
+ available_gpus: []
440
+ available_gpus: []
441
+ available_gpus: []
442
+ available_gpus: []
443
+ available_gpus: []
444
+ available_gpus: []
445
+ available_gpus: []
446
+ available_gpus: []
447
+ available_gpus: []
448
+ available_gpus: []
449
+ available_gpus: []
450
+ available_gpus: []
451
+ available_gpus: []
452
+ available_gpus: []
453
+ available_gpus: []
454
+ available_gpus: []
455
+ available_gpus: []
456
+ available_gpus: []
457
+ available_gpus: []
458
+ available_gpus: []
459
+ available_gpus: []
460
+ available_gpus: []
461
+ available_gpus: []
462
+ available_gpus: []
463
+ available_gpus: []
464
+ available_gpus: []
465
+ available_gpus: []
466
+ available_gpus: []
467
+ available_gpus: []
468
+ available_gpus: []
469
+ available_gpus: []
470
+ available_gpus: []
471
+ available_gpus: []
472
+ available_gpus: []
473
+ available_gpus: []
474
+ available_gpus: []
475
+ available_gpus: []
476
+ available_gpus: []
477
+ available_gpus: []
478
+ available_gpus: []
479
+ available_gpus: []
480
+ available_gpus: []
481
+ available_gpus: [2, 3]
482
+ The following command is about to run:
483
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,3 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/27/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-27 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/27/gaia/inf.log 2>&1 &
484
+ available_gpus: []
485
+ available_gpus: []
486
+ available_gpus: []
487
+ available_gpus: []
488
+ available_gpus: []
489
+ available_gpus: []
490
+ available_gpus: []
491
+ available_gpus: []
492
+ available_gpus: []
493
+ available_gpus: []
494
+ available_gpus: []
495
+ available_gpus: []
496
+ available_gpus: []
497
+ available_gpus: []
498
+ available_gpus: []
499
+ available_gpus: []
500
+ available_gpus: []
501
+ available_gpus: []
502
+ available_gpus: []
503
+ available_gpus: []
504
+ available_gpus: []
505
+ available_gpus: []
506
+ available_gpus: []
507
+ available_gpus: []
508
+ available_gpus: []
509
+ available_gpus: []
510
+ available_gpus: []
511
+ available_gpus: []
512
+ available_gpus: []
513
+ available_gpus: []
514
+ available_gpus: []
515
+ available_gpus: []
516
+ available_gpus: []
517
+ available_gpus: []
518
+ available_gpus: []
519
+ available_gpus: []
520
+ available_gpus: []
521
+ available_gpus: []
522
+ available_gpus: []
523
+ available_gpus: []
524
+ available_gpus: []
525
+ available_gpus: []
526
+ available_gpus: []
527
+ available_gpus: []
528
+ available_gpus: []
529
+ available_gpus: []
530
+ available_gpus: []
531
+ available_gpus: []
532
+ available_gpus: []
533
+ available_gpus: []
534
+ available_gpus: []
535
+ available_gpus: []
536
+ available_gpus: []
537
+ available_gpus: []
538
+ available_gpus: []
539
+ available_gpus: []
540
+ available_gpus: []
541
+ available_gpus: []
542
+ available_gpus: []
543
+ available_gpus: []
544
+ available_gpus: []
545
+ available_gpus: []
546
+ available_gpus: []
547
+ available_gpus: [2, 3]
548
+ The following command is about to run:
549
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,3 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name aime --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/27/aime --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-27 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/27/aime/inf.log 2>&1 &
550
+ available_gpus: []
551
+ available_gpus: []
552
+ available_gpus: []
553
+ available_gpus: []
554
+ available_gpus: []
555
+ available_gpus: []
556
+ available_gpus: []
557
+ available_gpus: []
558
+ available_gpus: [2, 3]
559
+ The following command is about to run:
560
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,3 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name frames --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/27/frames --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-27 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/27/frames/inf.log 2>&1 &
561
+ available_gpus: []
562
+ available_gpus: []
563
+ available_gpus: []
564
+ available_gpus: []
565
+ available_gpus: []
566
+ available_gpus: []
567
+ available_gpus: []
568
+ available_gpus: []
569
+ available_gpus: []
570
+ available_gpus: []
571
+ available_gpus: []
572
+ available_gpus: []
573
+ available_gpus: []
574
+ available_gpus: []
575
+ available_gpus: []
576
+ available_gpus: []
577
+ available_gpus: []
578
+ available_gpus: []
579
+ available_gpus: []
580
+ available_gpus: []
581
+ available_gpus: []
582
+ available_gpus: []
583
+ available_gpus: []
584
+ available_gpus: []
585
+ available_gpus: []
586
+ available_gpus: []
587
+ available_gpus: []
588
+ available_gpus: []
589
+ available_gpus: []
590
+ available_gpus: []
591
+ available_gpus: []
592
+ available_gpus: []
593
+ available_gpus: []
594
+ available_gpus: []
595
+ available_gpus: []
596
+ available_gpus: []
597
+ available_gpus: []
598
+ available_gpus: []
599
+ available_gpus: []
600
+ available_gpus: []
601
+ available_gpus: []
602
+ available_gpus: []
603
+ available_gpus: []
604
+ available_gpus: []
605
+ available_gpus: []
606
+ available_gpus: []
607
+ available_gpus: []
608
+ available_gpus: []
609
+ available_gpus: []
610
+ available_gpus: []
611
+ available_gpus: []
612
+ available_gpus: []
613
+ available_gpus: []
614
+ available_gpus: []
615
+ available_gpus: []
616
+ available_gpus: []
617
+ available_gpus: []
618
+ available_gpus: []
619
+ available_gpus: [4, 5]
620
+ The following command is about to run:
621
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name eval --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/13/eval --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-13 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/13/eval/inf.log 2>&1 &
622
+ available_gpus: []
623
+ available_gpus: [6, 7]
624
+ The following command is about to run:
625
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/13/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-13 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/13/gaia/inf.log 2>&1 &
626
+ available_gpus: []
627
+ available_gpus: []
628
+ available_gpus: []
629
+ available_gpus: []
630
+ available_gpus: []
631
+ available_gpus: []
632
+ available_gpus: []
633
+ available_gpus: []
634
+ available_gpus: []
635
+ available_gpus: []
636
+ available_gpus: []
637
+ available_gpus: []
638
+ available_gpus: []
639
+ available_gpus: []
640
+ available_gpus: []
641
+ available_gpus: []
642
+ available_gpus: []
643
+ available_gpus: []
644
+ available_gpus: []
645
+ available_gpus: []
646
+ available_gpus: []
647
+ available_gpus: []
648
+ available_gpus: []
649
+ available_gpus: []
650
+ available_gpus: []
651
+ available_gpus: []
652
+ available_gpus: []
653
+ available_gpus: []
654
+ available_gpus: []
655
+ available_gpus: []
656
+ available_gpus: []
657
+ available_gpus: []
658
+ available_gpus: []
659
+ available_gpus: []
660
+ available_gpus: []
661
+ available_gpus: []
662
+ available_gpus: []
663
+ available_gpus: []
664
+ available_gpus: []
665
+ available_gpus: []
666
+ available_gpus: []
667
+ available_gpus: []
668
+ available_gpus: []
669
+ available_gpus: []
670
+ available_gpus: []
671
+ available_gpus: []
672
+ available_gpus: []
673
+ available_gpus: []
674
+ available_gpus: []
675
+ available_gpus: []
676
+ available_gpus: []
677
+ available_gpus: []
678
+ available_gpus: []
679
+ available_gpus: []
680
+ available_gpus: []
681
+ available_gpus: []
682
+ available_gpus: []
683
+ available_gpus: []
684
+ available_gpus: []
685
+ available_gpus: []
686
+ available_gpus: []
687
+ available_gpus: []
688
+ available_gpus: []
689
+ available_gpus: []
690
+ available_gpus: []
691
+ available_gpus: []
692
+ available_gpus: []
693
+ available_gpus: []
694
+ available_gpus: []
695
+ available_gpus: []
696
+ available_gpus: []
697
+ available_gpus: []
698
+ available_gpus: []
699
+ available_gpus: []
700
+ available_gpus: []
701
+ available_gpus: []
702
+ available_gpus: []
703
+ available_gpus: []
704
+ available_gpus: []
705
+ available_gpus: []
706
+ available_gpus: []
707
+ available_gpus: []
708
+ available_gpus: []
709
+ available_gpus: []
710
+ available_gpus: []
711
+ available_gpus: []
712
+ available_gpus: []
713
+ available_gpus: []
714
+ available_gpus: []
715
+ available_gpus: []
716
+ available_gpus: []
717
+ available_gpus: []
718
+ available_gpus: []
719
+ available_gpus: []
720
+ available_gpus: []
721
+ available_gpus: []
722
+ available_gpus: []
723
+ available_gpus: []
724
+ available_gpus: []
725
+ available_gpus: []
726
+ available_gpus: []
727
+ available_gpus: []
728
+ available_gpus: []
729
+ available_gpus: []
730
+ available_gpus: []
731
+ available_gpus: [6, 7]
732
+ The following command is about to run:
733
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name aime --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/13/aime --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-13 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/13/aime/inf.log 2>&1 &
734
+ available_gpus: []
735
+ available_gpus: []
736
+ available_gpus: []
737
+ available_gpus: []
738
+ available_gpus: []
739
+ available_gpus: []
740
+ available_gpus: []
741
+ available_gpus: []
742
+ available_gpus: []
743
+ available_gpus: [6, 7]
744
+ The following command is about to run:
745
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name frames --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/13/frames --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/checkpoint-13 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math/13/frames/inf.log 2>&1 &
746
+ Wish me good luck!
deep_search/sft/4-4_no_error_data_871_doc_by_itself_DeepSeek-R1-Distill-Qwen-32.log ADDED
@@ -0,0 +1,115 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ nohup: ignoring input
2
+ step_list: [78, 68, 55, 41]
3
+ step_list: [78, 68, 55, 41]
4
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/checkpoint-78) to exist...
5
+ The checkpoint exists. Waiting for running...
6
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval_old_500 --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/78/eval_old_500 --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/checkpoint-78 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/78/eval_old_500/inf.log 2>&1 &
7
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/checkpoint-68) to exist...
8
+ The checkpoint exists. Waiting for running...
9
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval_old_500 --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/68/eval_old_500 --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/checkpoint-68 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/68/eval_old_500/inf.log 2>&1 &
10
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/checkpoint-55) to exist...
11
+ The checkpoint exists. Waiting for running...
12
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval_old_500 --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/55/eval_old_500 --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/checkpoint-55 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/55/eval_old_500/inf.log 2>&1 &
13
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/checkpoint-41) to exist...
14
+ The checkpoint exists. Waiting for running...
15
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval_old_500 --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/41/eval_old_500 --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/checkpoint-41 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/41/eval_old_500/inf.log 2>&1 &
16
+ All checkpoints exist. Wait for runing...
17
+ available_gpus: [6, 7]
18
+ The following command is about to run:
19
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval_old_500 --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/78/eval_old_500 --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/checkpoint-78 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/78/eval_old_500/inf.log 2>&1 &
20
+ available_gpus: []
21
+ available_gpus: []
22
+ available_gpus: []
23
+ available_gpus: []
24
+ available_gpus: []
25
+ available_gpus: []
26
+ available_gpus: []
27
+ available_gpus: []
28
+ available_gpus: []
29
+ available_gpus: []
30
+ available_gpus: []
31
+ available_gpus: []
32
+ available_gpus: []
33
+ available_gpus: []
34
+ available_gpus: []
35
+ available_gpus: []
36
+ available_gpus: []
37
+ available_gpus: []
38
+ available_gpus: []
39
+ available_gpus: [4, 5]
40
+ The following command is about to run:
41
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval_old_500 --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/68/eval_old_500 --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/checkpoint-68 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/68/eval_old_500/inf.log 2>&1 &
42
+ available_gpus: []
43
+ available_gpus: []
44
+ available_gpus: []
45
+ available_gpus: []
46
+ available_gpus: []
47
+ available_gpus: []
48
+ available_gpus: []
49
+ available_gpus: []
50
+ available_gpus: []
51
+ available_gpus: []
52
+ available_gpus: []
53
+ available_gpus: []
54
+ available_gpus: []
55
+ available_gpus: []
56
+ available_gpus: []
57
+ available_gpus: []
58
+ available_gpus: []
59
+ available_gpus: []
60
+ available_gpus: []
61
+ available_gpus: []
62
+ available_gpus: []
63
+ available_gpus: []
64
+ available_gpus: []
65
+ available_gpus: []
66
+ available_gpus: []
67
+ available_gpus: []
68
+ available_gpus: []
69
+ available_gpus: []
70
+ available_gpus: []
71
+ available_gpus: []
72
+ available_gpus: []
73
+ available_gpus: []
74
+ available_gpus: []
75
+ available_gpus: []
76
+ available_gpus: []
77
+ available_gpus: []
78
+ available_gpus: []
79
+ available_gpus: []
80
+ available_gpus: []
81
+ available_gpus: []
82
+ available_gpus: []
83
+ available_gpus: []
84
+ available_gpus: []
85
+ available_gpus: []
86
+ available_gpus: []
87
+ available_gpus: []
88
+ available_gpus: []
89
+ available_gpus: []
90
+ available_gpus: []
91
+ available_gpus: []
92
+ available_gpus: []
93
+ available_gpus: []
94
+ available_gpus: []
95
+ available_gpus: []
96
+ available_gpus: []
97
+ available_gpus: []
98
+ available_gpus: []
99
+ available_gpus: []
100
+ available_gpus: []
101
+ available_gpus: []
102
+ available_gpus: []
103
+ available_gpus: []
104
+ available_gpus: []
105
+ available_gpus: [2, 3]
106
+ The following command is about to run:
107
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,3 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval_old_500 --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/55/eval_old_500 --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/checkpoint-55 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/55/eval_old_500/inf.log 2>&1 &
108
+ available_gpus: []
109
+ available_gpus: []
110
+ available_gpus: []
111
+ available_gpus: []
112
+ available_gpus: [0, 1]
113
+ The following command is about to run:
114
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=0,1 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage.py --dataset_name eval_old_500 --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/41/eval_old_500 --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/checkpoint-41 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871/41/eval_old_500/inf.log 2>&1 &
115
+ Wish me good luck!
deep_search/sft/5-5_ablation_ques_domain_filtered_data_738_Qwen2.5-7B-Instruct.log ADDED
@@ -0,0 +1,128 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ step_list: [66, 57, 46, 34, 23]
2
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/checkpoint-66) to exist...
3
+ The checkpoint exists. Waiting for running...
4
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/66/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/checkpoint-66 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/66/gaia/inf.log 2>&1 &
5
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/66/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/checkpoint-66 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/66/bamboogle/inf.log 2>&1 &
6
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/checkpoint-57) to exist...
7
+ The checkpoint exists. Waiting for running...
8
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/57/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/checkpoint-57 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/57/gaia/inf.log 2>&1 &
9
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/57/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/checkpoint-57 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/57/bamboogle/inf.log 2>&1 &
10
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/checkpoint-46) to exist...
11
+ The checkpoint exists. Waiting for running...
12
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/46/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/checkpoint-46 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/46/gaia/inf.log 2>&1 &
13
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/46/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/checkpoint-46 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/46/bamboogle/inf.log 2>&1 &
14
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/checkpoint-34) to exist...
15
+ The checkpoint exists. Waiting for running...
16
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/34/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/checkpoint-34 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/34/gaia/inf.log 2>&1 &
17
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/34/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/checkpoint-34 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/34/bamboogle/inf.log 2>&1 &
18
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/checkpoint-23) to exist...
19
+ The checkpoint exists. Waiting for running...
20
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/23/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/checkpoint-23 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/23/gaia/inf.log 2>&1 &
21
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/23/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/checkpoint-23 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/23/bamboogle/inf.log 2>&1 &
22
+ All checkpoints exist. Wait for runing...
23
+ available_gpus: [2, 3, 4, 5, 6, 7]
24
+ The following command is about to run:
25
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,3 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/66/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/checkpoint-66 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/66/gaia/inf.log 2>&1 &
26
+ The following command is about to run:
27
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/66/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/checkpoint-66 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/66/bamboogle/inf.log 2>&1 &
28
+ The following command is about to run:
29
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/57/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/checkpoint-57 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/57/gaia/inf.log 2>&1 &
30
+ available_gpus: []
31
+ available_gpus: []
32
+ available_gpus: []
33
+ available_gpus: []
34
+ available_gpus: []
35
+ available_gpus: []
36
+ available_gpus: []
37
+ available_gpus: []
38
+ available_gpus: []
39
+ available_gpus: []
40
+ available_gpus: []
41
+ available_gpus: []
42
+ available_gpus: []
43
+ available_gpus: []
44
+ available_gpus: []
45
+ available_gpus: []
46
+ available_gpus: []
47
+ available_gpus: []
48
+ available_gpus: []
49
+ available_gpus: []
50
+ available_gpus: []
51
+ available_gpus: []
52
+ available_gpus: []
53
+ available_gpus: []
54
+ available_gpus: []
55
+ available_gpus: [4, 5]
56
+ The following command is about to run:
57
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/57/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/checkpoint-57 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/57/bamboogle/inf.log 2>&1 &
58
+ available_gpus: []
59
+ available_gpus: []
60
+ available_gpus: []
61
+ available_gpus: []
62
+ available_gpus: []
63
+ available_gpus: []
64
+ available_gpus: []
65
+ available_gpus: []
66
+ available_gpus: []
67
+ available_gpus: []
68
+ available_gpus: []
69
+ available_gpus: []
70
+ available_gpus: []
71
+ available_gpus: []
72
+ available_gpus: [4, 5]
73
+ The following command is about to run:
74
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/46/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/checkpoint-46 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/46/gaia/inf.log 2>&1 &
75
+ available_gpus: []
76
+ available_gpus: []
77
+ available_gpus: []
78
+ available_gpus: []
79
+ available_gpus: []
80
+ available_gpus: []
81
+ available_gpus: [6, 7]
82
+ The following command is about to run:
83
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/46/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/checkpoint-46 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/46/bamboogle/inf.log 2>&1 &
84
+ available_gpus: []
85
+ available_gpus: []
86
+ available_gpus: []
87
+ available_gpus: [2, 3]
88
+ The following command is about to run:
89
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,3 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/34/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/checkpoint-34 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/34/gaia/inf.log 2>&1 &
90
+ available_gpus: []
91
+ available_gpus: []
92
+ available_gpus: []
93
+ available_gpus: []
94
+ available_gpus: []
95
+ available_gpus: []
96
+ available_gpus: []
97
+ available_gpus: []
98
+ available_gpus: []
99
+ available_gpus: []
100
+ available_gpus: []
101
+ available_gpus: [6, 7]
102
+ The following command is about to run:
103
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/34/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/checkpoint-34 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/34/bamboogle/inf.log 2>&1 &
104
+ available_gpus: []
105
+ available_gpus: []
106
+ available_gpus: []
107
+ available_gpus: []
108
+ available_gpus: []
109
+ available_gpus: []
110
+ available_gpus: []
111
+ available_gpus: []
112
+ available_gpus: []
113
+ available_gpus: []
114
+ available_gpus: []
115
+ available_gpus: []
116
+ available_gpus: []
117
+ available_gpus: []
118
+ available_gpus: []
119
+ available_gpus: []
120
+ available_gpus: []
121
+ available_gpus: []
122
+ available_gpus: [6, 7]
123
+ The following command is about to run:
124
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/23/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/checkpoint-23 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/23/gaia/inf.log 2>&1 &
125
+ available_gpus: [4, 5]
126
+ The following command is about to run:
127
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/23/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/checkpoint-23 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:14123#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_domain_filtered_data_738_ablation/23/bamboogle/inf.log 2>&1 &
128
+ Wish me good luck!
deep_search/sft/5-5_ablation_ques_keywords_filtered_data_727_Qwen2.5-7B-Instruct.log ADDED
@@ -0,0 +1,128 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ step_list: [66, 57, 46, 34, 23]
2
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/checkpoint-66) to exist...
3
+ The checkpoint exists. Waiting for running...
4
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/66/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/checkpoint-66 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/66/gaia/inf.log 2>&1 &
5
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/66/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/checkpoint-66 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/66/bamboogle/inf.log 2>&1 &
6
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/checkpoint-57) to exist...
7
+ The checkpoint exists. Waiting for running...
8
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/57/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/checkpoint-57 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/57/gaia/inf.log 2>&1 &
9
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/57/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/checkpoint-57 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/57/bamboogle/inf.log 2>&1 &
10
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/checkpoint-46) to exist...
11
+ The checkpoint exists. Waiting for running...
12
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/46/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/checkpoint-46 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/46/gaia/inf.log 2>&1 &
13
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/46/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/checkpoint-46 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/46/bamboogle/inf.log 2>&1 &
14
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/checkpoint-34) to exist...
15
+ The checkpoint exists. Waiting for running...
16
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/34/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/checkpoint-34 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/34/gaia/inf.log 2>&1 &
17
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/34/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/checkpoint-34 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/34/bamboogle/inf.log 2>&1 &
18
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/checkpoint-23) to exist...
19
+ The checkpoint exists. Waiting for running...
20
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/23/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/checkpoint-23 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/23/gaia/inf.log 2>&1 &
21
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/23/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/checkpoint-23 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/23/bamboogle/inf.log 2>&1 &
22
+ All checkpoints exist. Wait for runing...
23
+ available_gpus: []
24
+ available_gpus: []
25
+ available_gpus: [6, 7]
26
+ The following command is about to run:
27
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/66/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/checkpoint-66 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/66/gaia/inf.log 2>&1 &
28
+ available_gpus: []
29
+ available_gpus: []
30
+ available_gpus: []
31
+ available_gpus: []
32
+ available_gpus: []
33
+ available_gpus: []
34
+ available_gpus: []
35
+ available_gpus: []
36
+ available_gpus: []
37
+ available_gpus: []
38
+ available_gpus: []
39
+ available_gpus: []
40
+ available_gpus: []
41
+ available_gpus: []
42
+ available_gpus: []
43
+ available_gpus: []
44
+ available_gpus: []
45
+ available_gpus: []
46
+ available_gpus: [4, 5, 6, 7]
47
+ The following command is about to run:
48
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/66/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/checkpoint-66 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/66/bamboogle/inf.log 2>&1 &
49
+ The following command is about to run:
50
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/57/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/checkpoint-57 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/57/gaia/inf.log 2>&1 &
51
+ available_gpus: [4, 5]
52
+ The following command is about to run:
53
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/57/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/checkpoint-57 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/57/bamboogle/inf.log 2>&1 &
54
+ available_gpus: [5]
55
+ available_gpus: [4, 5]
56
+ The following command is about to run:
57
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/46/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/checkpoint-46 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/46/gaia/inf.log 2>&1 &
58
+ available_gpus: []
59
+ available_gpus: []
60
+ available_gpus: [2, 3]
61
+ The following command is about to run:
62
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,3 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/46/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/checkpoint-46 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/46/bamboogle/inf.log 2>&1 &
63
+ available_gpus: []
64
+ available_gpus: []
65
+ available_gpus: []
66
+ available_gpus: []
67
+ available_gpus: []
68
+ available_gpus: []
69
+ available_gpus: []
70
+ available_gpus: []
71
+ available_gpus: []
72
+ available_gpus: []
73
+ available_gpus: []
74
+ available_gpus: []
75
+ available_gpus: []
76
+ available_gpus: []
77
+ available_gpus: []
78
+ available_gpus: []
79
+ available_gpus: []
80
+ available_gpus: []
81
+ available_gpus: []
82
+ available_gpus: []
83
+ available_gpus: []
84
+ available_gpus: []
85
+ available_gpus: []
86
+ available_gpus: []
87
+ available_gpus: [2, 3]
88
+ The following command is about to run:
89
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,3 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/34/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/checkpoint-34 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/34/gaia/inf.log 2>&1 &
90
+ available_gpus: []
91
+ available_gpus: []
92
+ available_gpus: []
93
+ available_gpus: []
94
+ available_gpus: []
95
+ available_gpus: []
96
+ available_gpus: []
97
+ available_gpus: []
98
+ available_gpus: []
99
+ available_gpus: [6, 7]
100
+ The following command is about to run:
101
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/34/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/checkpoint-34 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/34/bamboogle/inf.log 2>&1 &
102
+ available_gpus: []
103
+ available_gpus: []
104
+ available_gpus: []
105
+ available_gpus: []
106
+ available_gpus: []
107
+ available_gpus: []
108
+ available_gpus: []
109
+ available_gpus: []
110
+ available_gpus: []
111
+ available_gpus: []
112
+ available_gpus: []
113
+ available_gpus: []
114
+ available_gpus: []
115
+ available_gpus: [4, 5]
116
+ The following command is about to run:
117
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/23/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/checkpoint-23 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/23/gaia/inf.log 2>&1 &
118
+ available_gpus: []
119
+ available_gpus: []
120
+ available_gpus: []
121
+ available_gpus: []
122
+ available_gpus: []
123
+ available_gpus: []
124
+ available_gpus: []
125
+ available_gpus: [6, 7]
126
+ The following command is about to run:
127
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/23/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/checkpoint-23 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:29477#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_keywords_filtered_data_727_ablation/23/bamboogle/inf.log 2>&1 &
128
+ Wish me good luck!
deep_search/sft/5-5_ablation_ques_yiwenci_filtered_data_811_Qwen2.5-7B-Instruct.log ADDED
@@ -0,0 +1,173 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ step_list: [72, 63, 51, 38, 25]
2
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/checkpoint-72) to exist...
3
+ The checkpoint exists. Waiting for running...
4
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/72/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/checkpoint-72 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/72/gaia/inf.log 2>&1 &
5
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/72/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/checkpoint-72 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/72/bamboogle/inf.log 2>&1 &
6
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/checkpoint-63) to exist...
7
+ The checkpoint exists. Waiting for running...
8
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/63/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/checkpoint-63 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/63/gaia/inf.log 2>&1 &
9
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/63/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/checkpoint-63 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/63/bamboogle/inf.log 2>&1 &
10
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/checkpoint-51) to exist...
11
+ The checkpoint exists. Waiting for running...
12
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/51/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/checkpoint-51 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/51/gaia/inf.log 2>&1 &
13
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/51/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/checkpoint-51 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/51/bamboogle/inf.log 2>&1 &
14
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/checkpoint-38) to exist...
15
+ The checkpoint exists. Waiting for running...
16
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/38/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/checkpoint-38 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/38/gaia/inf.log 2>&1 &
17
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/38/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/checkpoint-38 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/38/bamboogle/inf.log 2>&1 &
18
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/checkpoint-25) to exist...
19
+ The checkpoint exists. Waiting for running...
20
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/25/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/checkpoint-25 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/25/gaia/inf.log 2>&1 &
21
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/25/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/checkpoint-25 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/25/bamboogle/inf.log 2>&1 &
22
+ All checkpoints exist. Wait for runing...
23
+ available_gpus: [6, 7]
24
+ The following command is about to run:
25
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/72/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/checkpoint-72 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/72/gaia/inf.log 2>&1 &
26
+ available_gpus: []
27
+ available_gpus: []
28
+ available_gpus: []
29
+ available_gpus: []
30
+ available_gpus: []
31
+ available_gpus: []
32
+ available_gpus: []
33
+ available_gpus: []
34
+ available_gpus: []
35
+ available_gpus: []
36
+ available_gpus: []
37
+ available_gpus: []
38
+ available_gpus: []
39
+ available_gpus: [4, 5]
40
+ The following command is about to run:
41
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/72/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/checkpoint-72 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/72/bamboogle/inf.log 2>&1 &
42
+ available_gpus: []
43
+ available_gpus: []
44
+ available_gpus: []
45
+ available_gpus: []
46
+ available_gpus: []
47
+ available_gpus: []
48
+ available_gpus: []
49
+ available_gpus: [6, 7]
50
+ The following command is about to run:
51
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/63/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/checkpoint-63 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/63/gaia/inf.log 2>&1 &
52
+ available_gpus: []
53
+ available_gpus: [6]
54
+ available_gpus: [6]
55
+ available_gpus: [6]
56
+ available_gpus: [6]
57
+ available_gpus: [6]
58
+ available_gpus: [6]
59
+ available_gpus: [6]
60
+ available_gpus: [6]
61
+ available_gpus: [6]
62
+ available_gpus: [6]
63
+ available_gpus: [2, 3, 6]
64
+ The following command is about to run:
65
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,3 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/63/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/checkpoint-63 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/63/bamboogle/inf.log 2>&1 &
66
+ available_gpus: [6]
67
+ available_gpus: [6]
68
+ available_gpus: [6]
69
+ available_gpus: [6]
70
+ available_gpus: [6]
71
+ available_gpus: [6]
72
+ available_gpus: [6]
73
+ available_gpus: [6]
74
+ available_gpus: [6]
75
+ available_gpus: [6]
76
+ available_gpus: [6]
77
+ available_gpus: [6]
78
+ available_gpus: [6]
79
+ available_gpus: [6]
80
+ available_gpus: [6]
81
+ available_gpus: [2, 3, 6]
82
+ The following command is about to run:
83
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,3 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/51/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/checkpoint-51 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/51/gaia/inf.log 2>&1 &
84
+ available_gpus: [6]
85
+ available_gpus: [6]
86
+ available_gpus: [6]
87
+ available_gpus: [6]
88
+ available_gpus: [6]
89
+ available_gpus: [6]
90
+ available_gpus: [6]
91
+ available_gpus: [6]
92
+ available_gpus: [6]
93
+ available_gpus: [6]
94
+ available_gpus: [6]
95
+ available_gpus: [6]
96
+ available_gpus: [6]
97
+ available_gpus: [6]
98
+ available_gpus: [6]
99
+ available_gpus: [6]
100
+ available_gpus: [6]
101
+ available_gpus: [6]
102
+ available_gpus: [6]
103
+ available_gpus: [6]
104
+ available_gpus: [6]
105
+ available_gpus: [6]
106
+ available_gpus: [6]
107
+ available_gpus: [6]
108
+ available_gpus: [6]
109
+ available_gpus: [4, 5, 6]
110
+ The following command is about to run:
111
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/51/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/checkpoint-51 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/51/bamboogle/inf.log 2>&1 &
112
+ available_gpus: [6]
113
+ available_gpus: [6]
114
+ available_gpus: [6]
115
+ available_gpus: [6]
116
+ available_gpus: [6]
117
+ available_gpus: [6]
118
+ available_gpus: [6]
119
+ available_gpus: [6]
120
+ available_gpus: [6]
121
+ available_gpus: [6]
122
+ available_gpus: [2, 3, 6]
123
+ The following command is about to run:
124
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,3 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/38/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/checkpoint-38 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/38/gaia/inf.log 2>&1 &
125
+ available_gpus: [6]
126
+ available_gpus: [6]
127
+ available_gpus: [6]
128
+ available_gpus: [6]
129
+ available_gpus: [6]
130
+ available_gpus: [4, 5, 6]
131
+ The following command is about to run:
132
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/38/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/checkpoint-38 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/38/bamboogle/inf.log 2>&1 &
133
+ available_gpus: [6]
134
+ available_gpus: [6]
135
+ available_gpus: [6]
136
+ available_gpus: [6]
137
+ available_gpus: [6]
138
+ available_gpus: [6]
139
+ available_gpus: [6]
140
+ available_gpus: [6]
141
+ available_gpus: [6]
142
+ available_gpus: [6]
143
+ available_gpus: [6]
144
+ available_gpus: [6]
145
+ available_gpus: [6]
146
+ available_gpus: [6]
147
+ available_gpus: [6]
148
+ available_gpus: [6]
149
+ available_gpus: [6]
150
+ available_gpus: [6]
151
+ available_gpus: [6]
152
+ available_gpus: [5, 6]
153
+ The following command is about to run:
154
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=5,6 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/25/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/checkpoint-25 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/25/gaia/inf.log 2>&1 &
155
+ available_gpus: [4]
156
+ available_gpus: [4]
157
+ available_gpus: [4]
158
+ available_gpus: [4]
159
+ available_gpus: [4]
160
+ available_gpus: [4]
161
+ available_gpus: [4]
162
+ available_gpus: [4]
163
+ available_gpus: [4]
164
+ available_gpus: [4]
165
+ available_gpus: [4]
166
+ available_gpus: [4]
167
+ available_gpus: [4]
168
+ available_gpus: [4]
169
+ available_gpus: [4]
170
+ available_gpus: [2, 3, 4]
171
+ The following command is about to run:
172
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,3 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/25/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/checkpoint-25 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:10634#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_ques_yiwenci_filtered_data_811_ablation/25/bamboogle/inf.log 2>&1 &
173
+ Wish me good luck!
deep_search/sft/5-5_ablation_resp_format_1064_random_sample_871_Qwen2.5-7B-Instruct.log ADDED
@@ -0,0 +1,72 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ step_list: [78, 68, 55, 41, 27]
2
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/checkpoint-78) to exist...
3
+ The checkpoint exists. Waiting for running...
4
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/78/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/checkpoint-78 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/78/gaia/inf.log 2>&1 &
5
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/78/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/checkpoint-78 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/78/bamboogle/inf.log 2>&1 &
6
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/checkpoint-68) to exist...
7
+ The checkpoint exists. Waiting for running...
8
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/68/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/checkpoint-68 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/68/gaia/inf.log 2>&1 &
9
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/68/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/checkpoint-68 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/68/bamboogle/inf.log 2>&1 &
10
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/checkpoint-55) to exist...
11
+ The checkpoint exists. Waiting for running...
12
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/55/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/checkpoint-55 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/55/gaia/inf.log 2>&1 &
13
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/55/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/checkpoint-55 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/55/bamboogle/inf.log 2>&1 &
14
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/checkpoint-41) to exist...
15
+ The checkpoint exists. Waiting for running...
16
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/41/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/checkpoint-41 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/41/gaia/inf.log 2>&1 &
17
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/41/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/checkpoint-41 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/41/bamboogle/inf.log 2>&1 &
18
+ Waiting for checkpoint (/capacity/userdata/ss/sft_search/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/checkpoint-27) to exist...
19
+ The checkpoint exists. Waiting for running...
20
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/27/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/checkpoint-27 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/27/gaia/inf.log 2>&1 &
21
+ command: export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES={device} nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/27/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/checkpoint-27 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/27/bamboogle/inf.log 2>&1 &
22
+ All checkpoints exist. Wait for runing...
23
+ available_gpus: []
24
+ available_gpus: []
25
+ available_gpus: []
26
+ available_gpus: []
27
+ available_gpus: []
28
+ available_gpus: []
29
+ available_gpus: []
30
+ available_gpus: []
31
+ available_gpus: []
32
+ available_gpus: []
33
+ available_gpus: []
34
+ available_gpus: []
35
+ available_gpus: []
36
+ available_gpus: []
37
+ available_gpus: []
38
+ available_gpus: []
39
+ available_gpus: []
40
+ available_gpus: []
41
+ available_gpus: [6, 7]
42
+ The following command is about to run:
43
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/78/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/checkpoint-78 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/78/gaia/inf.log 2>&1 &
44
+ available_gpus: []
45
+ available_gpus: [6, 7]
46
+ The following command is about to run:
47
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=6,7 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/78/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/checkpoint-78 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/78/bamboogle/inf.log 2>&1 &
48
+ available_gpus: []
49
+ available_gpus: [2, 3]
50
+ The following command is about to run:
51
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=2,3 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name gaia --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/68/gaia --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/checkpoint-68 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/68/gaia/inf.log 2>&1 &
52
+ available_gpus: []
53
+ available_gpus: []
54
+ available_gpus: []
55
+ available_gpus: [4, 5]
56
+ The following command is about to run:
57
+ export https_proxy=http://127.0.0.1:7890 && export http_proxy=http://127.0.0.1:7890 && export all_proxy=socks5://127.0.0.1:7891 && cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && CUDA_VISIBLE_DEVICES=4,5 nohup /opt/aps/workdir/search_o1/bin/python3 -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py --dataset_name bamboogle --cache_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/cache_eval_sum_all_webpage_1w --output_dir_base /opt/aps/workdir/sunshuang/deep_search/search_o1/output/output_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/68/bamboogle --split test --max_search_limit 10 --max_turn 10 --top_k 10 --max_doc_len 5000 --model_path /capacity/userdata/ss/sft_search/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/checkpoint-68 --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" > /opt/aps/workdir/sunshuang/deep_search/search_o1/logs/log_eval/JOB:30702#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_resp_format_1064_random_sample_871_ablation/68/bamboogle/inf.log 2>&1 &
58
+ available_gpus: []
59
+ available_gpus: []
60
+ available_gpus: []
61
+ available_gpus: []
62
+ available_gpus: []
63
+ available_gpus: []
64
+ available_gpus: []
65
+ available_gpus: []
66
+ available_gpus: []
67
+ available_gpus: []
68
+ available_gpus: []
69
+ available_gpus: []
70
+ available_gpus: []
71
+ available_gpus: []
72
+ available_gpus: []
deep_search/sft/ds_zero3_offload.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bf16": {
3
+ "enabled": "auto"
4
+ },
5
+ "zero_optimization": {
6
+ "stage": 3,
7
+ "offload_optimizer": {
8
+ "device": "cpu",
9
+ "pin_memory": true
10
+ },
11
+ "offload_param": {
12
+ "device": "cpu",
13
+ "pin_memory": true
14
+ },
15
+ "overlap_comm": true,
16
+ "contiguous_gradients": true,
17
+ "sub_group_size": 5E8,
18
+ "reduce_bucket_size": "auto",
19
+ "stage3_prefetch_bucket_size": "auto",
20
+ "stage3_param_persistence_threshold": "auto",
21
+ "stage3_max_live_parameters": 5E8,
22
+ "stage3_max_reuse_distance": 5E8,
23
+ "stage3_gather_16bit_weights_on_model_save": true
24
+ },
25
+ "gradient_accumulation_steps": "auto",
26
+ "gradient_clipping": "auto",
27
+ "steps_per_print": 2000,
28
+ "train_batch_size": "auto",
29
+ "train_micro_batch_size_per_gpu": "auto",
30
+ "wall_clock_breakdown": false,
31
+ "flops_profiler": {
32
+ "enabled": false,
33
+ "profile_step": 10,
34
+ "module_depth": -1,
35
+ "top_modules": 3,
36
+ "detailed": true,
37
+ "output_file": "flops_profiler.out"
38
+ }
39
+ }
deep_search/sft/gen_data_2.log ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ Traceback (most recent call last):
2
+ File "/opt/aps/workdir/sunshuang/deep_search/search_o1/gen_data.py", line 4, in <module>
3
+ from vllm import LLM, SamplingParams
4
+ ModuleNotFoundError: No module named 'vllm'
deep_search/sft/hostfile ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ lmlabide-824b9f7c-6967-4403-8cf6-8911048f903a-worker-0 slots=8
2
+ lmlabide-824b9f7c-6967-4403-8cf6-8911048f903a-worker-1 slots=8
deep_search/sft/mix.sh ADDED
@@ -0,0 +1,265 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/bin/bash
2
+
3
+ export OMP_NUM_THREADS=20
4
+ export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
5
+ # export CUDA_VISIBLE_DEVICES=0,1,2,3
6
+ # --include localhost:0,1,2,3,4,5,6,7 \
7
+
8
+ # /opt/aps/workdir/miniforge3/envs/train/bin/deepspeed \
9
+ # --master_port=9944 \
10
+ # --include localhost:4,5,6,7 \
11
+ # sft.py \
12
+ # --deepspeed ds_zero3.json \
13
+ # --model_name_or_path /capacity/userdata/models/Qwen2.5-32B-Instruct \
14
+ # --tokenizer_name_or_path /capacity/userdata/models/DeepSeek-R1-Distill-Qwen-1.5B \
15
+ # --do_train \
16
+ # --save_safetensors true \
17
+ # --data_path /opt/aps/workdir/input/data/search-o1/merged_selected_data.json \
18
+ # --lr_scheduler_type cosine \
19
+ # --output_dir output/checkpoint/qwen_32B_test \
20
+ # --overwrite_output_dir \
21
+ # --warmup_ratio 0.03 \
22
+ # --gradient_checkpointing true \
23
+ # --per_device_train_batch_size 2 \
24
+ # --gradient_accumulation_steps 2 \
25
+ # --logging_steps 1 \
26
+ # --learning_rate 2e-5 \
27
+ # --num_train_epochs 2 \
28
+ # --save_steps 400 \
29
+ # --model_max_length 8192 \
30
+ # --save_total_limit 16 \
31
+ # --bf16 || exit 1
32
+
33
+ # /opt/aps/workdir/miniforge3/envs/train/bin/deepspeed \
34
+ # --master_port=9944 \
35
+ # sft.py \
36
+ # --deepspeed ds_zero3.json \
37
+ # --model_name_or_path /capacity/userdata/models/Qwen2.5-32B-Instruct \
38
+ # --tokenizer_name_or_path /capacity/userdata/models/DeepSeek-R1-Distill-Qwen-32 \
39
+ # --do_train \
40
+ # --save_safetensors true \
41
+ # --data_path /opt/aps/workdir/input/data/search-o1/hopotqa_1217.json \
42
+ # --lr_scheduler_type cosine \
43
+ # --output_dir output/checkpoint/qwen_2_5_32b_data_1217 \
44
+ # --overwrite_output_dir \
45
+ # --warmup_ratio 0.03 \
46
+ # --gradient_checkpointing true \
47
+ # --per_device_train_batch_size 1 \
48
+ # --gradient_accumulation_steps 4 \
49
+ # --logging_steps 1 \
50
+ # --learning_rate 2e-5 \
51
+ # --num_train_epochs 1 \
52
+ # --save_steps 400 \
53
+ # --model_max_length 8192 \
54
+ # --save_total_limit 16 \
55
+ # --bf16 || exit 1
56
+
57
+ # /opt/aps/workdir/miniforge3/envs/train/bin/deepspeed \
58
+ # --master_port=9944 \
59
+ # --include localhost:4,5,6,7 \
60
+ # sft.py \
61
+ # --deepspeed ds_zero3.json \
62
+ # --model_name_or_path /opt/aps/workdir/home/.cache/modelscope/hub/models/qwen/Qwen2.5-7B-Instruct \
63
+ # --tokenizer_name_or_path /capacity/userdata/models/DeepSeek-R1-Distill-Qwen-32 \
64
+ # --do_train \
65
+ # --save_safetensors true \
66
+ # --data_path /opt/aps/workdir/input/data/search-o1/hopotqa_1217.json \
67
+ # --lr_scheduler_type cosine \
68
+ # --output_dir output/checkpoint/qwen_7b_inst_data_1217 \
69
+ # --overwrite_output_dir \
70
+ # --warmup_ratio 0.03 \
71
+ # --gradient_checkpointing true \
72
+ # --per_device_train_batch_size 1 \
73
+ # --gradient_accumulation_steps 4 \
74
+ # --logging_steps 1 \
75
+ # --learning_rate 2e-5 \
76
+ # --num_train_epochs 1 \
77
+ # --save_steps 400 \
78
+ # --model_max_length 8192 \
79
+ # --save_total_limit 16 \
80
+ # --bf16 || exit 1
81
+
82
+
83
+ # qwen 7b 用自己的tokenizer
84
+ # /opt/aps/workdir/miniforge3/envs/train/bin/deepspeed \
85
+ # --master_port=9944 \
86
+ # --include localhost:6,7 \
87
+ # sft.py \
88
+ # --deepspeed ds_zero3.json \
89
+ # --model_name_or_path /opt/aps/workdir/home/.cache/modelscope/hub/models/qwen/Qwen2.5-7B-Instruct \
90
+ # --tokenizer_name_or_path /opt/aps/workdir/home/.cache/modelscope/hub/models/qwen/Qwen2.5-7B-Instruct \
91
+ # --do_train \
92
+ # --save_safetensors true \
93
+ # --data_path /opt/aps/workdir/input/data/search-o1/hopotqa_1217.json \
94
+ # --lr_scheduler_type cosine \
95
+ # --output_dir output/checkpoint/qwen_7b_original_tokenizer_inst_data_1217 \
96
+ # --overwrite_output_dir \
97
+ # --warmup_ratio 0.03 \
98
+ # --gradient_checkpointing true \
99
+ # --per_device_train_batch_size 1 \
100
+ # --gradient_accumulation_steps 4 \
101
+ # --logging_steps 1 \
102
+ # --learning_rate 2e-5 \
103
+ # --num_train_epochs 1 \
104
+ # --save_steps 400 \
105
+ # --model_max_length 8192 \
106
+ # --save_total_limit 16 \
107
+ # --bf16 || exit 1
108
+
109
+ # /opt/aps/workdir/miniforge3/envs/train/bin/deepspeed \
110
+ # --master_port=9944 \
111
+ # --include localhost:7 \
112
+ # sft_1.py \
113
+ # --deepspeed ds_zero3.json \
114
+ # --model_name_or_path /opt/aps/workdir/home/.cache/modelscope/hub/models/qwen/Qwen2.5-7B-Instruct \
115
+ # --tokenizer_name_or_path /capacity/userdata/models/DeepSeek-R1-Distill-Qwen-32 \
116
+ # --do_train \
117
+ # --save_safetensors true \
118
+ # --data_path /opt/aps/workdir/input/data/search-o1/hopotqa_1217.json \
119
+ # --lr_scheduler_type cosine \
120
+ # --output_dir output/checkpoint/qwen_7b_inst_data_1217 \
121
+ # --overwrite_output_dir \
122
+ # --warmup_ratio 0.03 \
123
+ # --gradient_checkpointing true \
124
+ # --per_device_train_batch_size 1 \
125
+ # --gradient_accumulation_steps 4 \
126
+ # --logging_steps 1 \
127
+ # --learning_rate 2e-5 \
128
+ # --num_train_epochs 1 \
129
+ # --save_steps 400 \
130
+ # --model_max_length 8192 \
131
+ # --save_total_limit 16 \
132
+ # --bf16 || exit 1
133
+
134
+
135
+
136
+ # qwen 7b 用自己的tokenizer
137
+ # /opt/aps/workdir/miniforge3/envs/train/bin/deepspeed \
138
+ # --master_port=9944 \
139
+ # --include localhost:6,7 \
140
+ # sft.py \
141
+ # --deepspeed ds_zero3.json \
142
+ # --model_name_or_path /opt/aps/workdir/home/.cache/modelscope/hub/models/qwen/Qwen2.5-7B-Instruct \
143
+ # --tokenizer_name_or_path /opt/aps/workdir/home/.cache/modelscope/hub/models/qwen/Qwen2.5-7B-Instruct \
144
+ # --do_train \
145
+ # --save_safetensors true \
146
+ # --data_path /opt/aps/workdir/input/data/search-o1/hopotqa_1217.json \
147
+ # --lr_scheduler_type cosine \
148
+ # --output_dir output/sft_use_original_tokenizer/qwen_7b_original_tokenizer_inst_data_1217_1 \
149
+ # --overwrite_output_dir \
150
+ # --warmup_ratio 0.03 \
151
+ # --gradient_checkpointing true \
152
+ # --per_device_train_batch_size 1 \
153
+ # --gradient_accumulation_steps 4 \
154
+ # --logging_steps 1 \
155
+ # --learning_rate 2e-5 \
156
+ # --num_train_epochs 1 \
157
+ # --model_max_length 8192 \
158
+ # --save_total_limit 16 \
159
+ # --bf16 || exit 1
160
+ # 定义参数
161
+ lr=1e-5
162
+ base=QwQ-32B
163
+ tokenizer=QwQ-32B
164
+ # train_data=hopotqa_1217.json
165
+ train_data=new_instruction_2k_sft
166
+ bsz=2
167
+ acc=4
168
+
169
+ # 生成随机 JOB-ID
170
+ JOB_ID=$(( RANDOM % 100000 )) # 生成一个 0 到 99999 的随机数
171
+ save_path="JOB:${JOB_ID}#LR:${lr}#BASE:${base}#TOKEN:${tokenizer}#BSZ:${bsz}#ACC:${acc}_${train_data}"
172
+
173
+ # 输出路径
174
+ output_dir="/share/project/sunshuang/deep_search/output/${save_path}"
175
+
176
+ output_dir_1=${output_dir}
177
+ model_name_1=${base}
178
+ # 创建输出目录
179
+ mkdir -p "$output_dir"
180
+
181
+ echo ${output_dir}
182
+
183
+ # 执行 deepspeed 命令
184
+ /share/project/miniconda/envs/ss_train/bin/deepspeed \
185
+ --master_port=9944 \
186
+ sft_1.py \
187
+ --deepspeed ds_zero3_offload.json \
188
+ --model_name_or_path "/share/project/zhipengchen/model/${base}" \
189
+ --tokenizer_name_or_path "/share/project/zhipengchen/model/${tokenizer}" \
190
+ --do_train \
191
+ --save_safetensors true \
192
+ --data_path "/share/project/sunshuang/deep_search/search_o1/sft_data/${train_data}.json" \
193
+ --lr_scheduler_type cosine \
194
+ --output_dir "$output_dir" \
195
+ --overwrite_output_dir \
196
+ --warmup_ratio 0.03 \
197
+ --gradient_checkpointing true \
198
+ --per_device_train_batch_size "$bsz" \
199
+ --gradient_accumulation_steps "$acc" \
200
+ --logging_steps 1 \
201
+ --learning_rate "$lr" \
202
+ --num_train_epochs 6 \
203
+ --save_strategy epoch \
204
+ --save_only_model true \
205
+ --model_max_length 30000 \
206
+ --save_total_limit 15 \
207
+ --bf16 || exit 1
208
+
209
+ # 3-15 model_max_length 25000 -> 30000
210
+ #################################################
211
+ lr=1e-5
212
+ base=DeepSeek-R1-Distill-Qwen-32B
213
+ tokenizer=DeepSeek-R1-Distill-Qwen-32B
214
+ # train_data=hopotqa_1217.json
215
+ train_data=new_instruction_2k_sft
216
+ bsz=2
217
+ acc=4
218
+
219
+ # 生成随机 JOB-ID
220
+ JOB_ID=$(( RANDOM % 100000 )) # 生成一个 0 到 99999 的随机数
221
+ save_path="JOB:${JOB_ID}#LR:${lr}#BASE:${base}#TOKEN:${tokenizer}#BSZ:${bsz}#ACC:${acc}_${train_data}"
222
+
223
+ # 输出路径
224
+ output_dir="/share/project/sunshuang/deep_search/output/${save_path}"
225
+
226
+ output_dir_2=${output_dir}
227
+ model_name_2=${base}
228
+
229
+ # 创建输出目录
230
+ mkdir -p "$output_dir"
231
+
232
+ echo ${output_dir}
233
+
234
+ # 执行 deepspeed 命令
235
+ /share/project/miniconda/envs/ss_train/bin/deepspeed \
236
+ --master_port=9944 \
237
+ sft_1.py \
238
+ --deepspeed ds_zero3_offload.json \
239
+ --model_name_or_path "/share/project/zhipengchen/model/${base}" \
240
+ --tokenizer_name_or_path "/share/project/zhipengchen/model/${tokenizer}" \
241
+ --do_train \
242
+ --save_safetensors true \
243
+ --data_path "/share/project/sunshuang/deep_search/search_o1/sft_data/${train_data}.json" \
244
+ --lr_scheduler_type cosine \
245
+ --output_dir "$output_dir" \
246
+ --overwrite_output_dir \
247
+ --warmup_ratio 0.03 \
248
+ --gradient_checkpointing true \
249
+ --per_device_train_batch_size "$bsz" \
250
+ --gradient_accumulation_steps "$acc" \
251
+ --logging_steps 1 \
252
+ --learning_rate "$lr" \
253
+ --num_train_epochs 6 \
254
+ --save_strategy epoch \
255
+ --save_only_model true \
256
+ --model_max_length 30000 \
257
+ --save_total_limit 15 \
258
+ --bf16 || exit 1
259
+
260
+
261
+ bash test.sh $output_dir_1 $model_name_1
262
+
263
+ bash test.sh $output_dir_2 $model_name_2
264
+
265
+
deep_search/sft/mix_2.sh ADDED
@@ -0,0 +1,295 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/bin/bash
2
+
3
+ export OMP_NUM_THREADS=20
4
+ export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
5
+
6
+ ###########################
7
+ # 定义参数
8
+ lr=1e-5
9
+ base=Qwen2.5-7B-Instruct
10
+ tokenizer=Qwen2.5-7B-Instruct
11
+ # train_data=hopotqa_1217.json
12
+ train_data=ablation_subquery_1073_random_sample_871
13
+ bsz=2
14
+ acc=4
15
+
16
+ # 生成随机 JOB-ID
17
+ JOB_ID=$(( RANDOM % 100000 )) # 生成一个 0 到 99999 的随机数
18
+ save_path="JOB:${JOB_ID}#LR:${lr}#BASE:${base}#TOKEN:${tokenizer}#BSZ:${bsz}#ACC:${acc}_${train_data}_ablation"
19
+
20
+ # 输出路径
21
+ output_dir="/capacity/userdata/ss/sft_search/${save_path}"
22
+
23
+ output_dir_1=${output_dir}
24
+ model_name_1=${base}
25
+ dataset_1=${train_data}
26
+ # 创建输出目录
27
+ mkdir -p "$output_dir"
28
+
29
+ echo ${output_dir}
30
+
31
+ # 执行 deepspeed 命令
32
+ /opt/aps/workdir/miniforge3/envs/ss_train/bin/deepspeed \
33
+ --master_port=9944 \
34
+ sft_1.py \
35
+ --deepspeed ds_zero3_offload.json \
36
+ --model_name_or_path "/capacity/userdata/models/${base}" \
37
+ --tokenizer_name_or_path "/capacity/userdata/models/${tokenizer}" \
38
+ --do_train \
39
+ --save_safetensors true \
40
+ --data_path "/opt/aps/workdir/sunshuang/deep_search/search_o1/sft_data/${train_data}.json" \
41
+ --lr_scheduler_type cosine \
42
+ --output_dir "$output_dir" \
43
+ --overwrite_output_dir \
44
+ --warmup_ratio 0.03 \
45
+ --gradient_checkpointing true \
46
+ --per_device_train_batch_size "$bsz" \
47
+ --gradient_accumulation_steps "$acc" \
48
+ --logging_steps 1 \
49
+ --learning_rate "$lr" \
50
+ --num_train_epochs 6 \
51
+ --save_strategy epoch \
52
+ --save_only_model true \
53
+ --model_max_length 30000 \
54
+ --save_total_limit 5 \
55
+ --bf16 || exit 1
56
+
57
+
58
+ ########################### 2
59
+ # 定义参数
60
+ lr=1e-5
61
+ base=Qwen2.5-7B-Instruct
62
+ tokenizer=Qwen2.5-7B-Instruct
63
+ # train_data=hopotqa_1217.json
64
+ train_data=ablation_resp_format_1064_random_sample_871
65
+ bsz=2
66
+ acc=4
67
+
68
+ # 生成随机 JOB-ID
69
+ JOB_ID=$(( RANDOM % 100000 )) # 生成一个 0 到 99999 的随机数
70
+ save_path="JOB:${JOB_ID}#LR:${lr}#BASE:${base}#TOKEN:${tokenizer}#BSZ:${bsz}#ACC:${acc}_${train_data}_ablation"
71
+
72
+ # 输出路径
73
+ output_dir="/capacity/userdata/ss/sft_search/${save_path}"
74
+
75
+ output_dir_2=${output_dir}
76
+ model_name_2=${base}
77
+ dataset_2=${train_data}
78
+ # 创建输出目录
79
+ mkdir -p "$output_dir"
80
+
81
+ echo ${output_dir}
82
+
83
+ # 执行 deepspeed 命令
84
+ /opt/aps/workdir/miniforge3/envs/ss_train/bin/deepspeed \
85
+ --master_port=9944 \
86
+ sft_1.py \
87
+ --deepspeed ds_zero3_offload.json \
88
+ --model_name_or_path "/capacity/userdata/models/${base}" \
89
+ --tokenizer_name_or_path "/capacity/userdata/models/${tokenizer}" \
90
+ --do_train \
91
+ --save_safetensors true \
92
+ --data_path "/opt/aps/workdir/sunshuang/deep_search/search_o1/sft_data/${train_data}.json" \
93
+ --lr_scheduler_type cosine \
94
+ --output_dir "$output_dir" \
95
+ --overwrite_output_dir \
96
+ --warmup_ratio 0.03 \
97
+ --gradient_checkpointing true \
98
+ --per_device_train_batch_size "$bsz" \
99
+ --gradient_accumulation_steps "$acc" \
100
+ --logging_steps 1 \
101
+ --learning_rate "$lr" \
102
+ --num_train_epochs 6 \
103
+ --save_strategy epoch \
104
+ --save_only_model true \
105
+ --model_max_length 30000 \
106
+ --save_total_limit 5 \
107
+ --bf16 || exit 1
108
+
109
+ ########################### 3
110
+ # 定义参数
111
+ lr=1e-5
112
+ base=Qwen2.5-7B-Instruct
113
+ tokenizer=Qwen2.5-7B-Instruct
114
+ # train_data=hopotqa_1217.json
115
+ train_data=ablation_ques_domain_filtered_data_738
116
+ bsz=2
117
+ acc=4
118
+
119
+ # 生成随机 JOB-ID
120
+ JOB_ID=$(( RANDOM % 100000 )) # 生成一个 0 到 99999 的随机数
121
+ save_path="JOB:${JOB_ID}#LR:${lr}#BASE:${base}#TOKEN:${tokenizer}#BSZ:${bsz}#ACC:${acc}_${train_data}_ablation"
122
+
123
+ # 输出路径
124
+ output_dir="/capacity/userdata/ss/sft_search/${save_path}"
125
+
126
+ output_dir_3=${output_dir}
127
+ model_name_3=${base}
128
+ dataset_3=${train_data}
129
+ # 创建输出目录
130
+ mkdir -p "$output_dir"
131
+
132
+ echo ${output_dir}
133
+
134
+ # 执行 deepspeed 命令
135
+ /opt/aps/workdir/miniforge3/envs/ss_train/bin/deepspeed \
136
+ --master_port=9944 \
137
+ sft_1.py \
138
+ --deepspeed ds_zero3_offload.json \
139
+ --model_name_or_path "/capacity/userdata/models/${base}" \
140
+ --tokenizer_name_or_path "/capacity/userdata/models/${tokenizer}" \
141
+ --do_train \
142
+ --save_safetensors true \
143
+ --data_path "/opt/aps/workdir/sunshuang/deep_search/search_o1/sft_data/${train_data}.json" \
144
+ --lr_scheduler_type cosine \
145
+ --output_dir "$output_dir" \
146
+ --overwrite_output_dir \
147
+ --warmup_ratio 0.03 \
148
+ --gradient_checkpointing true \
149
+ --per_device_train_batch_size "$bsz" \
150
+ --gradient_accumulation_steps "$acc" \
151
+ --logging_steps 1 \
152
+ --learning_rate "$lr" \
153
+ --num_train_epochs 6 \
154
+ --save_strategy epoch \
155
+ --save_only_model true \
156
+ --model_max_length 30000 \
157
+ --save_total_limit 5 \
158
+ --bf16 || exit 1
159
+
160
+
161
+
162
+
163
+ ########################### 4
164
+ # 定义参数
165
+ lr=1e-5
166
+ base=Qwen2.5-7B-Instruct
167
+ tokenizer=Qwen2.5-7B-Instruct
168
+ # train_data=hopotqa_1217.json
169
+ train_data=ablation_ques_keywords_filtered_data_727
170
+ bsz=2
171
+ acc=4
172
+
173
+ # 生成随机 JOB-ID
174
+ JOB_ID=$(( RANDOM % 100000 )) # 生成一个 0 到 99999 的随机数
175
+ save_path="JOB:${JOB_ID}#LR:${lr}#BASE:${base}#TOKEN:${tokenizer}#BSZ:${bsz}#ACC:${acc}_${train_data}_ablation"
176
+
177
+ # 输出路径
178
+ output_dir="/capacity/userdata/ss/sft_search/${save_path}"
179
+
180
+ output_dir_4=${output_dir}
181
+ model_name_4=${base}
182
+ dataset_4=${train_data}
183
+ # 创建输出目录
184
+ mkdir -p "$output_dir"
185
+
186
+ echo ${output_dir}
187
+
188
+ # 执行 deepspeed 命令
189
+ /opt/aps/workdir/miniforge3/envs/ss_train/bin/deepspeed \
190
+ --master_port=9944 \
191
+ sft_1.py \
192
+ --deepspeed ds_zero3_offload.json \
193
+ --model_name_or_path "/capacity/userdata/models/${base}" \
194
+ --tokenizer_name_or_path "/capacity/userdata/models/${tokenizer}" \
195
+ --do_train \
196
+ --save_safetensors true \
197
+ --data_path "/opt/aps/workdir/sunshuang/deep_search/search_o1/sft_data/${train_data}.json" \
198
+ --lr_scheduler_type cosine \
199
+ --output_dir "$output_dir" \
200
+ --overwrite_output_dir \
201
+ --warmup_ratio 0.03 \
202
+ --gradient_checkpointing true \
203
+ --per_device_train_batch_size "$bsz" \
204
+ --gradient_accumulation_steps "$acc" \
205
+ --logging_steps 1 \
206
+ --learning_rate "$lr" \
207
+ --num_train_epochs 6 \
208
+ --save_strategy epoch \
209
+ --save_only_model true \
210
+ --model_max_length 30000 \
211
+ --save_total_limit 5 \
212
+ --bf16 || exit 1
213
+
214
+
215
+
216
+
217
+ ########################### 5
218
+ # 定义参数
219
+ lr=1e-5
220
+ base=Qwen2.5-7B-Instruct
221
+ tokenizer=Qwen2.5-7B-Instruct
222
+ # train_data=hopotqa_1217.json
223
+ train_data=ablation_ques_yiwenci_filtered_data_811
224
+ bsz=2
225
+ acc=4
226
+
227
+ # 生成随机 JOB-ID
228
+ JOB_ID=$(( RANDOM % 100000 )) # 生成一个 0 到 99999 的随机数
229
+ save_path="JOB:${JOB_ID}#LR:${lr}#BASE:${base}#TOKEN:${tokenizer}#BSZ:${bsz}#ACC:${acc}_${train_data}_ablation"
230
+
231
+ # 输出路径
232
+ output_dir="/capacity/userdata/ss/sft_search/${save_path}"
233
+
234
+ output_dir_5=${output_dir}
235
+ model_name_5=${base}
236
+ dataset_5=${train_data}
237
+ # 创建输出目录
238
+ mkdir -p "$output_dir"
239
+
240
+ echo ${output_dir}
241
+
242
+ # 执行 deepspeed 命令
243
+ /opt/aps/workdir/miniforge3/envs/ss_train/bin/deepspeed \
244
+ --master_port=9944 \
245
+ sft_1.py \
246
+ --deepspeed ds_zero3_offload.json \
247
+ --model_name_or_path "/capacity/userdata/models/${base}" \
248
+ --tokenizer_name_or_path "/capacity/userdata/models/${tokenizer}" \
249
+ --do_train \
250
+ --save_safetensors true \
251
+ --data_path "/opt/aps/workdir/sunshuang/deep_search/search_o1/sft_data/${train_data}.json" \
252
+ --lr_scheduler_type cosine \
253
+ --output_dir "$output_dir" \
254
+ --overwrite_output_dir \
255
+ --warmup_ratio 0.03 \
256
+ --gradient_checkpointing true \
257
+ --per_device_train_batch_size "$bsz" \
258
+ --gradient_accumulation_steps "$acc" \
259
+ --logging_steps 1 \
260
+ --learning_rate "$lr" \
261
+ --num_train_epochs 6 \
262
+ --save_strategy epoch \
263
+ --save_only_model true \
264
+ --model_max_length 30000 \
265
+ --save_total_limit 5 \
266
+ --bf16 || exit 1
267
+
268
+
269
+ ###################################
270
+ # 测试
271
+
272
+
273
+ bash test_two_model_qwq.sh $output_dir_1 $model_name_1 $dataset_1
274
+
275
+ sleep 7200
276
+ bash test_two_model_qwq_1.sh $output_dir_2 $model_name_2 $dataset_2
277
+
278
+ sleep 7200
279
+ bash test_two_model_qwq_1.sh $output_dir_3 $model_name_3 $dataset_3
280
+
281
+ sleep 7200
282
+ bash test_two_model_qwq_1.sh $output_dir_4 $model_name_4 $dataset_4
283
+
284
+ sleep 7200
285
+ bash test_two_model_qwq_1.sh $output_dir_5 $model_name_5 $dataset_5
286
+
287
+
288
+ nohup python /opt/aps/workdir/sunshuang/deep_search/search_o1/gen_data.py \
289
+ --input_file_path "/opt/aps/workdir/sunshuang/deep_search/data_syn/data/mixed_data/mixed_data_all.json" \
290
+ --cuda_visible_devices "0,1,2,3" > gen_data_1.log 2>&1 &
291
+
292
+
293
+ nohup python /opt/aps/workdir/sunshuang/deep_search/search_o1/gen_data.py \
294
+ --input_file_path "/opt/aps/workdir/sunshuang/deep_search/data_syn/data/mixed_data/mixed_data_all.json" \
295
+ --cuda_visible_devices "4,5,6,7" > gen_data_2.log 2>&1 &
deep_search/sft/mix_2_1.sh ADDED
@@ -0,0 +1,174 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/bin/bash
2
+
3
+ export OMP_NUM_THREADS=20
4
+ export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
5
+
6
+ # ###########################
7
+ # # 定义参数
8
+ # lr=1e-5
9
+ # base=Qwen2.5-7B-Instruct
10
+ # tokenizer=Qwen2.5-7B-Instruct
11
+ # # train_data=hopotqa_1217.json
12
+ # train_data=ablation_resp_difficulty_1678_random_sample_871
13
+ # bsz=2
14
+ # acc=4
15
+
16
+ # # 生成随机 JOB-ID
17
+ # JOB_ID=$(( RANDOM % 100000 )) # 生成一个 0 到 99999 的随机数
18
+ # save_path="JOB:${JOB_ID}#LR:${lr}#BASE:${base}#TOKEN:${tokenizer}#BSZ:${bsz}#ACC:${acc}_${train_data}_ablation"
19
+
20
+ # # 输出路径
21
+ # output_dir="/capacity/userdata/ss/sft_search/${save_path}"
22
+
23
+ # output_dir_1=${output_dir}
24
+ # model_name_1=${base}
25
+ # dataset_1=${train_data}
26
+ # # 创建输出目录
27
+ # mkdir -p "$output_dir"
28
+
29
+ # echo ${output_dir}
30
+
31
+ # # 执行 deepspeed 命令
32
+ # /opt/aps/workdir/miniforge3/envs/ss_train/bin/deepspeed \
33
+ # --master_port=9944 \
34
+ # sft_1.py \
35
+ # --deepspeed ds_zero3_offload.json \
36
+ # --model_name_or_path "/capacity/userdata/models/${base}" \
37
+ # --tokenizer_name_or_path "/capacity/userdata/models/${tokenizer}" \
38
+ # --do_train \
39
+ # --save_safetensors true \
40
+ # --data_path "/opt/aps/workdir/sunshuang/deep_search/search_o1/sft_data/${train_data}.json" \
41
+ # --lr_scheduler_type cosine \
42
+ # --output_dir "$output_dir" \
43
+ # --overwrite_output_dir \
44
+ # --warmup_ratio 0.03 \
45
+ # --gradient_checkpointing true \
46
+ # --per_device_train_batch_size "$bsz" \
47
+ # --gradient_accumulation_steps "$acc" \
48
+ # --logging_steps 1 \
49
+ # --learning_rate "$lr" \
50
+ # --num_train_epochs 6 \
51
+ # --save_strategy epoch \
52
+ # --save_only_model true \
53
+ # --model_max_length 30000 \
54
+ # --save_total_limit 2 \
55
+ # --bf16 || exit 1
56
+
57
+
58
+ ########################### 2
59
+ # 定义参数
60
+ lr=1e-5
61
+ base=Qwen2.5-7B-Instruct
62
+ tokenizer=Qwen2.5-7B-Instruct
63
+ # train_data=hopotqa_1217.json
64
+ train_data=ablation_resp_format_1064_random_sample_871
65
+ bsz=2
66
+ acc=4
67
+
68
+ # 生成随机 JOB-ID
69
+ JOB_ID=$(( RANDOM % 100000 )) # 生成一个 0 到 99999 的随机数
70
+ save_path="JOB:${JOB_ID}#LR:${lr}#BASE:${base}#TOKEN:${tokenizer}#BSZ:${bsz}#ACC:${acc}_${train_data}_ablation"
71
+
72
+ # 输出路径
73
+ output_dir="/capacity/userdata/ss/sft_search/${save_path}"
74
+
75
+ output_dir_2=${output_dir}
76
+ model_name_2=${base}
77
+ dataset_2=${train_data}
78
+ # 创建输出目录
79
+ mkdir -p "$output_dir"
80
+
81
+ echo ${output_dir}
82
+
83
+ # 执行 deepspeed 命令
84
+ /opt/aps/workdir/miniforge3/envs/ss_train/bin/deepspeed \
85
+ --master_port=9944 \
86
+ sft_1.py \
87
+ --deepspeed ds_zero3_offload.json \
88
+ --model_name_or_path "/capacity/userdata/models/${base}" \
89
+ --tokenizer_name_or_path "/capacity/userdata/models/${tokenizer}" \
90
+ --do_train \
91
+ --save_safetensors true \
92
+ --data_path "/opt/aps/workdir/sunshuang/deep_search/search_o1/sft_data/${train_data}.json" \
93
+ --lr_scheduler_type cosine \
94
+ --output_dir "$output_dir" \
95
+ --overwrite_output_dir \
96
+ --warmup_ratio 0.03 \
97
+ --gradient_checkpointing true \
98
+ --per_device_train_batch_size "$bsz" \
99
+ --gradient_accumulation_steps "$acc" \
100
+ --logging_steps 1 \
101
+ --learning_rate "$lr" \
102
+ --num_train_epochs 6 \
103
+ --save_strategy epoch \
104
+ --save_only_model true \
105
+ --model_max_length 30000 \
106
+ --save_total_limit 2 \
107
+ --bf16 || exit 1
108
+
109
+
110
+
111
+ # #### 3
112
+ # lr=1e-5
113
+ # base=Qwen2.5-7B-Instruct
114
+ # tokenizer=Qwen2.5-7B-Instruct
115
+ # # train_data=hopotqa_1217.json
116
+ # train_data=ablation_subquery_1073
117
+ # bsz=2
118
+ # acc=4
119
+
120
+ # # 生成随机 JOB-ID
121
+ # JOB_ID=$(( RANDOM % 100000 )) # 生成一个 0 到 99999 的随机数
122
+ # save_path="JOB:${JOB_ID}#LR:${lr}#BASE:${base}#TOKEN:${tokenizer}#BSZ:${bsz}#ACC:${acc}_${train_data}_ablation"
123
+
124
+ # # 输出路径
125
+ # output_dir="/capacity/userdata/ss/sft_search/${save_path}"
126
+
127
+ # output_dir_1=${output_dir}
128
+ # model_name_1=${base}
129
+ # dataset_1=${train_data}
130
+ # # 创建输出目录
131
+ # mkdir -p "$output_dir"
132
+
133
+ # echo ${output_dir}
134
+
135
+ # # 执行 deepspeed 命令
136
+ # /opt/aps/workdir/miniforge3/envs/ss_train/bin/deepspeed \
137
+ # --master_port=9944 \
138
+ # sft_1.py \
139
+ # --deepspeed ds_zero3_offload.json \
140
+ # --model_name_or_path "/capacity/userdata/models/${base}" \
141
+ # --tokenizer_name_or_path "/capacity/userdata/models/${tokenizer}" \
142
+ # --do_train \
143
+ # --save_safetensors true \
144
+ # --data_path "/opt/aps/workdir/sunshuang/deep_search/search_o1/sft_data/${train_data}.json" \
145
+ # --lr_scheduler_type cosine \
146
+ # --output_dir "$output_dir" \
147
+ # --overwrite_output_dir \
148
+ # --warmup_ratio 0.03 \
149
+ # --gradient_checkpointing true \
150
+ # --per_device_train_batch_size "$bsz" \
151
+ # --gradient_accumulation_steps "$acc" \
152
+ # --logging_steps 1 \
153
+ # --learning_rate "$lr" \
154
+ # --num_train_epochs 6 \
155
+ # --save_strategy epoch \
156
+ # --save_only_model true \
157
+ # --model_max_length 30000 \
158
+ # --save_total_limit 2 \
159
+ # --bf16 || exit 1
160
+
161
+ ###################################
162
+ # 测试
163
+
164
+
165
+ # bash test_two_model_qwq.sh $output_dir_1 $model_name_1 $dataset_1
166
+
167
+ # sleep 7200
168
+ # bash test_two_model_qwq.sh $output_dir_2 $model_name_2 $dataset_2
169
+
170
+ # sleep 600
171
+ # bash test.sh $output_dir_3 $model_name_3 $dataset_3
172
+
173
+
174
+
deep_search/sft/mix_math.sh ADDED
@@ -0,0 +1,62 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/bin/bash
2
+
3
+ export OMP_NUM_THREADS=20
4
+ export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
5
+ # 定义参数
6
+ lr=1e-5
7
+ base=QwQ-32B
8
+ tokenizer=QwQ-32B
9
+ # train_data=hopotqa_1217.json
10
+ train_data=no_error_data_871
11
+ bsz=2
12
+ acc=4
13
+
14
+ # 生成随机 JOB-ID
15
+ JOB_ID=$(( RANDOM % 100000 )) # 生成一个 0 到 99999 的随机数
16
+ save_path="JOB:${JOB_ID}#LR:${lr}#BASE:${base}#TOKEN:${tokenizer}#BSZ:${bsz}#ACC:${acc}_${train_data}_mixed_math"
17
+
18
+ # 输出路径
19
+ output_dir="/capacity/userdata/ss/sft_search/${save_path}"
20
+
21
+ output_dir_1=${output_dir}
22
+ model_name_1=${base}
23
+ dataset_1=${train_data}
24
+ # 创建输出目录
25
+ mkdir -p "$output_dir"
26
+
27
+ echo ${output_dir}
28
+
29
+ # 执行 deepspeed 命令
30
+ /opt/aps/workdir/miniforge3/envs/ss_train/bin/deepspeed \
31
+ --master_port=9944 \
32
+ sft_2_math.py \
33
+ --deepspeed ds_zero3_offload.json \
34
+ --model_name_or_path "/capacity/userdata/models/${base}" \
35
+ --tokenizer_name_or_path "/capacity/userdata/models/${tokenizer}" \
36
+ --do_train \
37
+ --save_safetensors true \
38
+ --data_path "/opt/aps/workdir/sunshuang/deep_search/search_o1/sft_data/${train_data}.json" \
39
+ --lr_scheduler_type cosine \
40
+ --output_dir "$output_dir" \
41
+ --overwrite_output_dir \
42
+ --warmup_ratio 0.03 \
43
+ --gradient_checkpointing true \
44
+ --per_device_train_batch_size "$bsz" \
45
+ --gradient_accumulation_steps "$acc" \
46
+ --logging_steps 1 \
47
+ --learning_rate "$lr" \
48
+ --num_train_epochs 1 \
49
+ --save_strategy epoch \
50
+ --save_only_model true \
51
+ --model_max_length 30000 \
52
+ --save_total_limit 4 \
53
+ --bf16 || exit 1
54
+
55
+
56
+ # bash test_two_model_qwq.sh $output_dir_1 $model_name_1 $dataset_1
57
+ # bash test.sh $output_dir_2 $model_name_2
58
+ # bash test.sh $output_dir_3 $model_name_3
59
+
60
+ # bash test.sh $output_dir_2 $model_name_2
61
+
62
+
deep_search/sft/mix_math_after_search.sh ADDED
@@ -0,0 +1,61 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/bin/bash
2
+
3
+ export OMP_NUM_THREADS=20
4
+ export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
5
+ # 定义参数
6
+ lr=1e-5
7
+ base=qwen7b_sft_871_checkpoint-78
8
+ tokenizer=qwen7b_sft_871_checkpoint-78
9
+ # train_data=hopotqa_1217.json
10
+ train_data=math_qwq_4524_selected_add_prompt_871
11
+ bsz=2
12
+ acc=4
13
+
14
+ # 生成随机 JOB-ID
15
+ JOB_ID=$(( RANDOM % 100000 )) # 生成一个 0 到 99999 的随机数
16
+ save_path="JOB:${JOB_ID}#LR:${lr}#BASE:${base}#TOKEN:${tokenizer}#BSZ:${bsz}#ACC:${acc}_${train_data}"
17
+
18
+ # 输出路径
19
+ output_dir="/capacity/userdata/ss/sft_search/${save_path}"
20
+
21
+ output_dir_1=${output_dir}
22
+ model_name_1=${base}
23
+ dataset_1=${train_data}
24
+ # 创建输出目录
25
+ mkdir -p "$output_dir"
26
+
27
+ echo ${output_dir}
28
+
29
+ # 执行 deepspeed 命令
30
+ /opt/aps/workdir/miniforge3/envs/ss_train/bin/deepspeed \
31
+ --master_port=9944 \
32
+ sft_2_math_after_search.py \
33
+ --deepspeed ds_zero3_offload.json \
34
+ --model_name_or_path "/capacity/userdata/ss/sft_search/qwen7b_sft_871_checkpoint-78" \
35
+ --tokenizer_name_or_path "/capacity/userdata/ss/sft_search/qwen7b_sft_871_checkpoint-78" \
36
+ --do_train \
37
+ --save_safetensors true \
38
+ --data_path "/opt/aps/workdir/sunshuang/deep_search/search_o1/sft_data/${train_data}.json" \
39
+ --lr_scheduler_type cosine \
40
+ --output_dir "$output_dir" \
41
+ --overwrite_output_dir \
42
+ --warmup_ratio 0.03 \
43
+ --gradient_checkpointing true \
44
+ --per_device_train_batch_size "$bsz" \
45
+ --gradient_accumulation_steps "$acc" \
46
+ --logging_steps 1 \
47
+ --learning_rate "$lr" \
48
+ --num_train_epochs 6 \
49
+ --save_strategy epoch \
50
+ --save_only_model true \
51
+ --model_max_length 30000 \
52
+ --save_total_limit 4 \
53
+ --bf16 || exit 1
54
+
55
+ bash test_two_model.sh $output_dir_1 $model_name_1 $dataset_1
56
+ # bash test.sh $output_dir_2 $model_name_2
57
+ # bash test.sh $output_dir_3 $model_name_3
58
+
59
+ # bash test.sh $output_dir_2 $model_name_2
60
+
61
+
deep_search/sft/mix_math_multi_node_1.sh ADDED
@@ -0,0 +1,66 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/bin/bash
2
+
3
+ export OMP_NUM_THREADS=20
4
+ export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
5
+ # 定义参数
6
+ lr=1e-5
7
+ base=QwQ-32B
8
+ tokenizer=QwQ-32B
9
+ # train_data=hopotqa_1217.json
10
+ train_data=no_error_data_871
11
+ bsz=1
12
+ acc=8
13
+
14
+ # 生成随机 JOB-ID
15
+ JOB_ID=$(( RANDOM % 100000 )) # 生成一个 0 到 99999 的随机数
16
+ save_path="JOB:${JOB_ID}#LR:${lr}#BASE:${base}#TOKEN:${tokenizer}#BSZ:${bsz}#ACC:${acc}_${train_data}_mixed_math"
17
+
18
+ # 输出路径
19
+ output_dir="/capacity/userdata/ss/sft_search/${save_path}"
20
+
21
+ output_dir_1=${output_dir}
22
+ model_name_1=${base}
23
+ dataset_1=${train_data}
24
+ # 创建输出目录
25
+ mkdir -p "$output_dir"
26
+
27
+ echo ${output_dir}
28
+
29
+ # 执行 deepspeed 命令
30
+ /opt/aps/workdir/miniforge3/envs/ss_train/bin/deepspeed \
31
+ --hostfile=hostfile \
32
+ --no_ssh \
33
+ --node_rank=1 \
34
+ --master_addr=172.19.164.116 \
35
+ --master_port=9944 \
36
+ sft_2_math.py \
37
+ --deepspeed ds_zero3_offload.json \
38
+ --model_name_or_path "/capacity/userdata/models/${base}" \
39
+ --tokenizer_name_or_path "/capacity/userdata/models/${tokenizer}" \
40
+ --do_train \
41
+ --save_safetensors true \
42
+ --data_path "/opt/aps/workdir/sunshuang/deep_search/search_o1/sft_data/${train_data}.json" \
43
+ --lr_scheduler_type cosine \
44
+ --output_dir "$output_dir" \
45
+ --overwrite_output_dir \
46
+ --warmup_ratio 0.03 \
47
+ --gradient_checkpointing true \
48
+ --per_device_train_batch_size "$bsz" \
49
+ --gradient_accumulation_steps "$acc" \
50
+ --logging_steps 1 \
51
+ --learning_rate "$lr" \
52
+ --num_train_epochs 6 \
53
+ --save_strategy epoch \
54
+ --save_only_model true \
55
+ --model_max_length 30000 \
56
+ --save_total_limit 6 \
57
+ --bf16 || exit 1
58
+
59
+
60
+ bash test_two_model_qwq_1.sh $output_dir_1 $model_name_1 $dataset_1
61
+ # bash test.sh $output_dir_2 $model_name_2
62
+ # bash test.sh $output_dir_3 $model_name_3
63
+
64
+ # bash test.sh $output_dir_2 $model_name_2
65
+
66
+
deep_search/sft/mix_math_sht_new_prompt.sh ADDED
@@ -0,0 +1,62 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/bin/bash
2
+
3
+ export OMP_NUM_THREADS=20
4
+ export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
5
+ # 定义参数
6
+ lr=1e-5
7
+ base=Qwen2.5-7B-Instruct
8
+ tokenizer=Qwen2.5-7B-Instruct
9
+ # train_data=hopotqa_1217.json
10
+ train_data=no_error_data_871_sht_new_prompt
11
+ bsz=2
12
+ acc=4
13
+
14
+ # 生成随机 JOB-ID
15
+ JOB_ID=$(( RANDOM % 100000 )) # 生成一个 0 到 99999 的随机数
16
+ save_path="JOB:${JOB_ID}#LR:${lr}#BASE:${base}#TOKEN:${tokenizer}#BSZ:${bsz}#ACC:${acc}_${train_data}_mixed_math_sht_new_prmopt"
17
+
18
+ # 输出路径
19
+ output_dir="/capacity/userdata/ss/sft_search/${save_path}"
20
+
21
+ output_dir_1=${output_dir}
22
+ model_name_1=${base}
23
+ dataset_1=${train_data}
24
+ # 创建输出目录
25
+ mkdir -p "$output_dir"
26
+
27
+ echo ${output_dir}
28
+
29
+ # 执行 deepspeed 命令
30
+ /opt/aps/workdir/miniforge3/envs/ss_train/bin/deepspeed \
31
+ --master_port=9944 \
32
+ sft_2_math_sht_new_prompt.py \
33
+ --deepspeed ds_zero3_offload.json \
34
+ --model_name_or_path "/capacity/userdata/models/${base}" \
35
+ --tokenizer_name_or_path "/capacity/userdata/models/${tokenizer}" \
36
+ --do_train \
37
+ --save_safetensors true \
38
+ --data_path "/opt/aps/workdir/sunshuang/deep_search/search_o1/sft_data/${train_data}.json" \
39
+ --lr_scheduler_type cosine \
40
+ --output_dir "$output_dir" \
41
+ --overwrite_output_dir \
42
+ --warmup_ratio 0.03 \
43
+ --gradient_checkpointing true \
44
+ --per_device_train_batch_size "$bsz" \
45
+ --gradient_accumulation_steps "$acc" \
46
+ --logging_steps 1 \
47
+ --learning_rate "$lr" \
48
+ --num_train_epochs 5 \
49
+ --save_strategy epoch \
50
+ --save_only_model true \
51
+ --model_max_length 30000 \
52
+ --save_total_limit 3 \
53
+ --bf16 || exit 1
54
+
55
+
56
+ # bash test_two_model_qwq.sh $output_dir_1 $model_name_1 $dataset_1
57
+ # bash test.sh $output_dir_2 $model_name_2
58
+ # bash test.sh $output_dir_3 $model_name_3
59
+
60
+ # bash test.sh $output_dir_2 $model_name_2
61
+
62
+
deep_search/sft/mix_wo_mask.sh ADDED
@@ -0,0 +1,62 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/bin/bash
2
+
3
+ export OMP_NUM_THREADS=20
4
+ export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
5
+ # 定义参数
6
+ lr=1e-5
7
+ base=Qwen2.5-7B-Instruct
8
+ tokenizer=Qwen2.5-7B-Instruct
9
+ # train_data=hopotqa_1217.json
10
+ train_data=no_error_data_871
11
+ bsz=2
12
+ acc=4
13
+
14
+ # 生成随机 JOB-ID
15
+ JOB_ID=$(( RANDOM % 100000 )) # 生成一个 0 到 99999 的随机数
16
+ save_path="JOB:${JOB_ID}#LR:${lr}#BASE:${base}#TOKEN:${tokenizer}#BSZ:${bsz}#ACC:${acc}_${train_data}_871_wo_mask"
17
+
18
+ # 输出路径
19
+ output_dir="/capacity/userdata/ss/sft_search/${save_path}"
20
+
21
+ output_dir_1=${output_dir}
22
+ model_name_1=${base}
23
+ dataset_1=${train_data}
24
+ # 创建输出目录
25
+ mkdir -p "$output_dir"
26
+
27
+ echo ${output_dir}
28
+
29
+ # 执行 deepspeed 命令
30
+ /opt/aps/workdir/miniforge3/envs/ss_train/bin/deepspeed \
31
+ --master_port=9944 \
32
+ sft_2_wo_mask.py \
33
+ --deepspeed ds_zero3_offload.json \
34
+ --model_name_or_path "/capacity/userdata/models/${base}" \
35
+ --tokenizer_name_or_path "/capacity/userdata/models/${tokenizer}" \
36
+ --do_train \
37
+ --save_safetensors true \
38
+ --data_path "/opt/aps/workdir/sunshuang/deep_search/search_o1/sft_data/${train_data}.json" \
39
+ --lr_scheduler_type cosine \
40
+ --output_dir "$output_dir" \
41
+ --overwrite_output_dir \
42
+ --warmup_ratio 0.03 \
43
+ --gradient_checkpointing true \
44
+ --per_device_train_batch_size "$bsz" \
45
+ --gradient_accumulation_steps "$acc" \
46
+ --logging_steps 1 \
47
+ --learning_rate "$lr" \
48
+ --num_train_epochs 6 \
49
+ --save_strategy epoch \
50
+ --save_only_model true \
51
+ --model_max_length 30000 \
52
+ --save_total_limit 6 \
53
+ --bf16 || exit 1
54
+
55
+
56
+ # bash test_two_model_qwq.sh $output_dir_1 $model_name_1 $dataset_1
57
+ # bash test.sh $output_dir_2 $model_name_2
58
+ # bash test.sh $output_dir_3 $model_name_3
59
+
60
+ # bash test.sh $output_dir_2 $model_name_2
61
+
62
+
deep_search/sft/run.sh ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ export HOME=/opt/aps/workdir/home
2
+ export http_proxy=http://127.0.0.1:7880
3
+ export https_proxy=http://127.0.0.1:7880
4
+
5
+ ./miniforge3/bin/conda init bash
6
+ source ~/.bashrc
7
+ eval "$(conda shell.bash hook)"
8
+ conda activate llm
9
+ bash mix.sh
10
+ # set -x
11
+
12
+ # bash Miniforge3.sh -b -u -p ./miniforge3
13
+ # ./miniforge3/bin/conda init bash
14
+ # source ~/.bashrc
15
+
16
+ # # export https_proxy='http://agent.baidu.com:8891'
17
+ # # export http_proxy='http://agent.baidu.com:8891'
18
+
19
+ # eval "$(conda shell.bash hook)"
20
+ # conda create -n llm python=3.11 -y
21
+ # conda activate llm
22
+
23
+ # pip install -r requirements.txt -i https://pypi.tuna.tsinghua.edu.cn/simple
24
+
25
+ cd /opt/aps/workdir/sunshuang/search_o1/ && http_proxy=http://127.0.0.1:7880 https_proxy=http://127.0.0.1:7880 CUDA_VISIBLE_DEVICES=2,3 /opt/aps/workdir/miniforge3/envs/search_o1/bin/python scripts/reason_two_model_2.py --dataset_name hotpotqa --cache_dir_base /opt/aps/workdir/sunshuang/search_o1/cache_reason_two_model/eval_reason_two_model/qwen-instruct-32B/JOB9986:LR1e-5:BASEQwen2.5-32B-Instruct:TOKENDeepSeek-R1-Distill-Qwen-32:BSZ1:ACC8/38/hotpotqa --output_dir_base /opt/aps/workdir/sunshuang/search_o1/outputs_reason_two_model/eval_reason_two_model/qwen-instruct-32B/JOB9986:LR1e-5:BASEQwen2.5-32B-Instruct:TOKENDeepSeek-R1-Distill-Qwen-32:BSZ1:ACC8/38/hotpotqa --split test --max_search_limit 5 --max_turn 10 --top_k 5 --max_doc_len 3000 --subset_num 100 --model_path /opt/aps/workdir/output/JOB9986:LR1e-5:BASEQwen2.5-32B-Instruct:TOKENDeepSeek-R1-Distill-Qwen-32:BSZ1:ACC8/checkpoint-38 --model_doc_reason_path "/capacity/userdata/models/Qwen2.5-32B-Instruct" --jina_api_key "jina_04d684ee4cc54d2ebe7c43bb7ad4aff0qlkdZGwm14NFBtm5BDkgK9KNf6vQ" --bing_subscription_key "cb0d28279a826d7e5cf22d71f683c77ffd4ba27d" --bing_endpoint "https://google.serper.dev/search" --openai_api_base "http://localhost:8001/v1"
deep_search/sft/sft.py ADDED
@@ -0,0 +1,204 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import os
2
+ import copy
3
+ import json
4
+ import logging
5
+ from tqdm import tqdm
6
+ from dataclasses import dataclass, field
7
+ from typing import Optional, Dict, Sequence
8
+
9
+ import torch
10
+ from torch.utils.data import random_split
11
+ from torch.nn.utils.rnn import pad_sequence
12
+ import transformers
13
+ from torch.utils.data import Dataset
14
+ from transformers import Trainer
15
+ import random
16
+ from typing import List, Optional, Tuple, Union
17
+ from transformers import AutoModelForCausalLM, TrainingArguments
18
+ from datasets import load_dataset
19
+ from transformers import DataCollatorForSeq2Seq
20
+ import shutil
21
+
22
+ # from liger_kernel.transformers import AutoLigerKernelForCausalLM
23
+
24
+
25
+ import matplotlib.pyplot as plt
26
+ import numpy as np
27
+
28
+
29
+ @dataclass
30
+ class ModelArguments:
31
+ model_name_or_path: Optional[str] = field(default="facebook/opt-125m")
32
+ # flash_attention: Optional[bool] = field(default=False)
33
+ tokenizer_name_or_path: Optional[str] = field(default=None)
34
+
35
+
36
+ @dataclass
37
+ class DataArguments:
38
+ data_path: str = field(
39
+ default=None, metadata={"help": "Path to the training data."}
40
+ )
41
+ prompt_type: Optional[str] = field(default="instruction")
42
+ dailog_augmentation: Optional[bool] = field(default=False)
43
+
44
+
45
+ @dataclass
46
+ class TrainingArguments(transformers.TrainingArguments):
47
+ cache_dir: Optional[str] = field(default=None)
48
+ optim: str = field(default="adamw_torch")
49
+ model_max_length: int = field(
50
+ default=512,
51
+ metadata={
52
+ "help": "Maximum sequence length. Sequences will be right padded (and possibly truncated)."
53
+ },
54
+ )
55
+
56
+
57
+ IGNORE_INDEX = -100
58
+ MAX_LENGTH = 2000
59
+
60
+ def process(sample, tokenizer):
61
+ # build inputs with format `<bos> X Y <eos>` and labels with format `<ignore> ... <ignore> Y <eos>`
62
+ # for multiturn examples, we only mask the prompt part in each prompt-response pair.
63
+ source = sample["input"]
64
+ # print(source)
65
+ # print(tokenizer.bos_token)
66
+ # print(source == None)
67
+ # if tokenizer.bos_token not in source:
68
+ # source = tokenizer.apply_chat_template(
69
+ # [
70
+ # {'role': 'user', 'content': source}
71
+ # ],
72
+ # tokenize=False, add_generation_prompt=True
73
+ # )
74
+ # # print(source)
75
+ source = tokenizer.apply_chat_template(
76
+ [
77
+ {'role': 'user', 'content': source}
78
+ ],
79
+ tokenize=False, add_generation_prompt=True
80
+ )
81
+ # print(source)
82
+ source = tokenizer(source, add_special_tokens=False)["input_ids"]
83
+ target = [IGNORE_INDEX] * len(source)
84
+ for output in sample["output"]:
85
+ for k, v in output.items():
86
+ if v is None:
87
+ continue
88
+ v_tokens = tokenizer(v, add_special_tokens=False)["input_ids"]
89
+ if k in ["gen"]:
90
+ source += v_tokens
91
+ target += v_tokens
92
+ elif k in ["doc_gen"]:
93
+ source += v_tokens
94
+ target += [IGNORE_INDEX] * len(v_tokens)
95
+ input_ids = source
96
+ labels = target
97
+
98
+ input_ids.append(tokenizer.eos_token_id)
99
+ labels.append(tokenizer.eos_token_id)
100
+ # if tokenizer.eos_token_id not in source:
101
+ # input_ids.append(tokenizer.eos_token_id)
102
+ # if tokenizer.eos_token_id not in labels:
103
+ # labels.append(tokenizer.eos_token_id)
104
+
105
+ # if len(input_ids) > MAX_LENGTH: # 删除过长的数据，防止爆显存
106
+ # return None
107
+ result = {
108
+ "input_ids": input_ids,
109
+ "attention_mask": [1] * len(input_ids),
110
+ "labels": labels,
111
+ }
112
+ # print(result)
113
+ return result
114
+
115
+
116
+ def print_function(example, tokenizer):
117
+ print("input_ids:\n{}".format(example["input_ids"]))
118
+ print("inputs:\n{}".format(tokenizer.decode(example["input_ids"], skip_special_tokens=False)))
119
+ print("label_ids:\n{}".format(example["labels"]))
120
+ print("labels:\n{}".format(
121
+ tokenizer.decode(list(filter(lambda x: x != IGNORE_INDEX, example["labels"])), skip_special_tokens=False)
122
+ ))
123
+
124
+
125
+ def get_dataset(file_path, tokenizer):
126
+ dataset = load_dataset('json', data_files=file_path)
127
+ train_dataset = dataset["train"]
128
+ file_name = os.path.basename(file_path)
129
+ dataset_name = os.path.splitext(file_name)[0]
130
+ # print(f"dataset_name: {dataset_name}")
131
+ # if os.path.exists(f"input/real_cache/{dataset_name}/"):
132
+ # shutil.rmtree(f"input/real_cache/{dataset_name}/")
133
+ # os.makedirs(f"input/real_cache/{dataset_name}/", exist_ok=True)
134
+ # tokenized_dataset = train_dataset.map(process, fn_kwargs={'tokenizer': tokenizer}, num_proc=1, cache_file_name=f"input/real_cache/{dataset_name}/cache1.arrow")
135
+ tokenized_dataset = train_dataset.map(process, fn_kwargs={'tokenizer': tokenizer}, num_proc=1, load_from_cache_file=False)
136
+ print_function(next(iter(tokenized_dataset)), tokenizer)
137
+ print(f"len of dataset before filter: {len(tokenized_dataset)}")
138
+
139
+ filtered_dataset = []
140
+ for item in tokenized_dataset:
141
+ if len(item["input_ids"]) <= 8000:
142
+ filtered_dataset.append(item)
143
+ print(f"len of dataset after filter: {len(filtered_dataset)}")
144
+ return filtered_dataset
145
+
146
+
147
+ def train():
148
+ parser = transformers.HfArgumentParser(
149
+ (ModelArguments, DataArguments, TrainingArguments)
150
+ )
151
+ model_args, data_args, training_args = parser.parse_args_into_dataclasses()
152
+
153
+ print("==========Model Args=========")
154
+ print(model_args)
155
+ print("==========Data Args=========")
156
+ print(data_args)
157
+ print("==========Training Args=========")
158
+ print(training_args)
159
+
160
+ if training_args.gradient_checkpointing:
161
+ use_cache = False # use_cache与gradient_checkpointing不能同时设置为true
162
+ model = AutoModelForCausalLM.from_pretrained(
163
+ model_args.model_name_or_path,
164
+ _attn_implementation="flash_attention_2",
165
+ use_cache=use_cache,
166
+ # save_only_model=True
167
+ ).float()
168
+ # model = AutoLigerKernelForCausalLM.from_pretrained(
169
+ # model_args.model_name_or_path,
170
+ # _attn_implementation="flash_attention_2",
171
+ # use_cache=use_cache,
172
+ # # save_only_model=True
173
+ # ).float()
174
+ if model_args.tokenizer_name_or_path is None:
175
+ model_args.tokenizer_name_or_path = model_args.model_name_or_path
176
+ tokenizer = transformers.AutoTokenizer.from_pretrained(
177
+ model_args.tokenizer_name_or_path, model_max_length=training_args.model_max_length
178
+ )
179
+
180
+ if tokenizer.pad_token is None:
181
+ tokenizer.pad_token = tokenizer.eos_token
182
+
183
+ dataset = get_dataset(data_args.data_path, tokenizer)
184
+
185
+ data_collator = DataCollatorForSeq2Seq(
186
+ tokenizer=tokenizer,
187
+ label_pad_token_id=-100,
188
+ )
189
+
190
+ trainer = Trainer(
191
+ model=model,
192
+ args=training_args,
193
+ tokenizer=tokenizer,
194
+ data_collator=data_collator,
195
+ train_dataset=dataset,
196
+ )
197
+ trainer.train(resume_from_checkpoint=training_args.resume_from_checkpoint)
198
+ trainer.save_model(training_args.output_dir)
199
+ trainer.save_state()
200
+
201
+
202
+ if __name__ == "__main__":
203
+ torch.manual_seed(42)
204
+ train()
deep_search/sft/sft_2_math_after_search.py ADDED
@@ -0,0 +1,260 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import os
2
+ import copy
3
+ import json
4
+ import logging
5
+ from tqdm import tqdm
6
+ from dataclasses import dataclass, field
7
+ from typing import Optional, Dict, Sequence
8
+
9
+ import torch
10
+ from torch.utils.data import random_split
11
+ from torch.nn.utils.rnn import pad_sequence
12
+ import transformers
13
+ from torch.utils.data import Dataset
14
+ from transformers import Trainer
15
+ import random
16
+ from typing import List, Optional, Tuple, Union
17
+ from transformers import AutoModelForCausalLM, TrainingArguments
18
+ from datasets import load_dataset
19
+ from transformers import DataCollatorForSeq2Seq
20
+ import shutil
21
+
22
+ # from liger_kernel.transformers import AutoLigerKernelForCausalLM
23
+
24
+
25
+ import matplotlib.pyplot as plt
26
+ import numpy as np
27
+
28
+
29
+ @dataclass
30
+ class ModelArguments:
31
+ model_name_or_path: Optional[str] = field(default="facebook/opt-125m")
32
+ # flash_attention: Optional[bool] = field(default=False)
33
+ tokenizer_name_or_path: Optional[str] = field(default=None)
34
+
35
+
36
+ @dataclass
37
+ class DataArguments:
38
+ data_path: str = field(
39
+ default=None, metadata={"help": "Path to the training data."}
40
+ )
41
+ prompt_type: Optional[str] = field(default="instruction")
42
+ dailog_augmentation: Optional[bool] = field(default=False)
43
+
44
+
45
+ @dataclass
46
+ class TrainingArguments(transformers.TrainingArguments):
47
+ cache_dir: Optional[str] = field(default=None)
48
+ optim: str = field(default="adamw_torch")
49
+ model_max_length: int = field(
50
+ default=512,
51
+ metadata={
52
+ "help": "Maximum sequence length. Sequences will be right padded (and possibly truncated)."
53
+ },
54
+ )
55
+
56
+
57
+ IGNORE_INDEX = -100
58
+ MAX_LENGTH = 2000
59
+
60
+ def process(sample, tokenizer):
61
+ # build inputs with format `<bos> X Y <eos>` and labels with format `<ignore> ... <ignore> Y <eos>`
62
+ # for multiturn examples, we only mask the prompt part in each prompt-response pair.
63
+ source = sample["input"]
64
+ # print(source)
65
+ # print(tokenizer.bos_token)
66
+ # print(source == None)
67
+ # if tokenizer.bos_token not in source:
68
+ # source = tokenizer.apply_chat_template(
69
+ # [
70
+ # {'role': 'user', 'content': source}
71
+ # ],
72
+ # tokenize=False, add_generation_prompt=True
73
+ # )
74
+ # # print(source)
75
+ source = tokenizer.apply_chat_template(
76
+ [
77
+ {'role': 'user', 'content': source}
78
+ ],
79
+ tokenize=False, add_generation_prompt=True
80
+ )
81
+ # print(source)
82
+ source = tokenizer(source, add_special_tokens=False)["input_ids"]
83
+ target = [IGNORE_INDEX] * len(source)
84
+ for output in sample["output"]:
85
+ for k, v in output.items():
86
+ if v is None:
87
+ continue
88
+ v_tokens = tokenizer(v, add_special_tokens=False)["input_ids"]
89
+ if k in ["gen"]:
90
+ source += v_tokens
91
+ target += v_tokens
92
+ elif k in ["doc_gen"]:
93
+ source += v_tokens
94
+ target += [IGNORE_INDEX] * len(v_tokens)
95
+ input_ids = source
96
+ labels = target
97
+
98
+ input_ids.append(tokenizer.eos_token_id)
99
+ labels.append(tokenizer.eos_token_id)
100
+ # if tokenizer.eos_token_id not in source:
101
+ # input_ids.append(tokenizer.eos_token_id)
102
+ # if tokenizer.eos_token_id not in labels:
103
+ # labels.append(tokenizer.eos_token_id)
104
+
105
+ # if len(input_ids) > MAX_LENGTH: # 删除过长的数据，防止爆显存
106
+ # return None
107
+ result = {
108
+ "input_ids": input_ids,
109
+ "attention_mask": [1] * len(input_ids),
110
+ "labels": labels,
111
+ }
112
+ # print(result)
113
+ return result
114
+
115
+
116
+ def process_math(sample, tokenizer):
117
+ # build inputs with format `<bos> X Y <eos>` and labels with format `<ignore> ... <ignore> Y <eos>`
118
+ # for multiturn examples, we only mask the prompt part in each prompt-response pair.
119
+ source = sample["prompt"]
120
+ source = tokenizer.apply_chat_template(
121
+ [
122
+ {'role': 'user', 'content': source}
123
+ ],
124
+ tokenize=False, add_generation_prompt=True
125
+ )
126
+
127
+ source = tokenizer(source, add_special_tokens=False)["input_ids"]
128
+ target = [IGNORE_INDEX] * len(source)
129
+
130
+ output = sample["output"]
131
+ output = tokenizer(output, add_special_tokens=False)["input_ids"]
132
+
133
+ source += output
134
+ target += output
135
+
136
+ input_ids = source
137
+ labels = target
138
+
139
+ input_ids.append(tokenizer.eos_token_id)
140
+ labels.append(tokenizer.eos_token_id)
141
+ # if tokenizer.eos_token_id not in source:
142
+ # input_ids.append(tokenizer.eos_token_id)
143
+ # if tokenizer.eos_token_id not in labels:
144
+ # labels.append(tokenizer.eos_token_id)
145
+
146
+ # if len(input_ids) > MAX_LENGTH: # 删除过长的数据，防止爆显存
147
+ # return None
148
+ result = {
149
+ "input_ids": input_ids,
150
+ "attention_mask": [1] * len(input_ids),
151
+ "labels": labels,
152
+ }
153
+ # print(result)
154
+ return result
155
+
156
+
157
+
158
+ def print_function(example, tokenizer):
159
+ print("input_ids:\n{}".format(example["input_ids"]))
160
+ print("inputs:\n{}".format(tokenizer.decode(example["input_ids"], skip_special_tokens=False)))
161
+ print("label_ids:\n{}".format(example["labels"]))
162
+ print("labels:\n{}".format(
163
+ tokenizer.decode(list(filter(lambda x: x != IGNORE_INDEX, example["labels"])), skip_special_tokens=False)
164
+ ))
165
+
166
+
167
+ def get_dataset(file_path, tokenizer, math_dataset=False):
168
+ dataset = load_dataset('json', data_files=file_path)
169
+ train_dataset = dataset["train"]
170
+ file_name = os.path.basename(file_path)
171
+ dataset_name = os.path.splitext(file_name)[0]
172
+ # print(f"dataset_name: {dataset_name}")
173
+ # if os.path.exists(f"input/real_cache/{dataset_name}/"):
174
+ # shutil.rmtree(f"input/real_cache/{dataset_name}/")
175
+ # os.makedirs(f"input/real_cache/{dataset_name}/", exist_ok=True)
176
+ # tokenized_dataset = train_dataset.map(process, fn_kwargs={'tokenizer': tokenizer}, num_proc=1, cache_file_name=f"input/real_cache/{dataset_name}/cache1.arrow")
177
+
178
+ if math_dataset:
179
+ tokenized_dataset = train_dataset.map(process_math, fn_kwargs={'tokenizer': tokenizer}, num_proc=1, load_from_cache_file=False)
180
+ else:
181
+ tokenized_dataset = train_dataset.map(process, fn_kwargs={'tokenizer': tokenizer}, num_proc=1, load_from_cache_file=False)
182
+ print_function(next(iter(tokenized_dataset)), tokenizer)
183
+ print(f"len of dataset before filter: {len(tokenized_dataset)}")
184
+
185
+ print(f"type of tokenized_dataset: {type(tokenized_dataset)}")
186
+
187
+ filtered_dataset = []
188
+ for item in tokenized_dataset:
189
+ # if len(item["input_ids"]) <= 10000:
190
+ # print(item)
191
+ print(f"type of item: {type(item)}")
192
+ filtered_dataset.append(item)
193
+ print(f"len of dataset after filter: {len(filtered_dataset)}")
194
+ return filtered_dataset
195
+
196
+
197
+ def train():
198
+ parser = transformers.HfArgumentParser(
199
+ (ModelArguments, DataArguments, TrainingArguments)
200
+ )
201
+ model_args, data_args, training_args = parser.parse_args_into_dataclasses()
202
+
203
+ print("==========Model Args=========")
204
+ print(model_args)
205
+ print("==========Data Args=========")
206
+ print(data_args)
207
+ print("==========Training Args=========")
208
+ print(training_args)
209
+
210
+ if training_args.gradient_checkpointing:
211
+ use_cache = False # use_cache与gradient_checkpointing不能同时设置为true
212
+ model = AutoModelForCausalLM.from_pretrained(
213
+ model_args.model_name_or_path,
214
+ _attn_implementation="flash_attention_2",
215
+ use_cache=use_cache,
216
+ # save_only_model=True
217
+ ).float()
218
+ # model = AutoLigerKernelForCausalLM.from_pretrained(
219
+ # model_args.model_name_or_path,
220
+ # _attn_implementation="flash_attention_2",
221
+ # use_cache=use_cache,
222
+ # # save_only_model=True
223
+ # ).float()
224
+ if model_args.tokenizer_name_or_path is None:
225
+ model_args.tokenizer_name_or_path = model_args.model_name_or_path
226
+ tokenizer = transformers.AutoTokenizer.from_pretrained(
227
+ model_args.tokenizer_name_or_path, model_max_length=training_args.model_max_length
228
+ )
229
+
230
+ if tokenizer.pad_token is None:
231
+ tokenizer.pad_token = tokenizer.eos_token
232
+
233
+ # dataset_qa = get_dataset(data_args.data_path, tokenizer)[:5]
234
+ # print(f"qa dataset: {len(dataset_qa)}")
235
+ dataset_math = get_dataset("/opt/aps/workdir/sunshuang/deep_search/search_o1/sft_data/math_qwq_4524_selected_add_prompt_871.json", tokenizer, True)
236
+ print(f"math dataset: {len(dataset_math)}")
237
+ dataset = dataset_math
238
+ # dataset = get_dataset(data_args.data_path, tokenizer)
239
+ print(f"fianl dataset: {len(dataset)}")
240
+
241
+ data_collator = DataCollatorForSeq2Seq(
242
+ tokenizer=tokenizer,
243
+ label_pad_token_id=-100,
244
+ )
245
+
246
+ trainer = Trainer(
247
+ model=model,
248
+ args=training_args,
249
+ tokenizer=tokenizer,
250
+ data_collator=data_collator,
251
+ train_dataset=dataset,
252
+ )
253
+ trainer.train(resume_from_checkpoint=training_args.resume_from_checkpoint)
254
+ trainer.save_model(training_args.output_dir)
255
+ trainer.save_state()
256
+
257
+
258
+ if __name__ == "__main__":
259
+ torch.manual_seed(42)
260
+ train()
deep_search/sft/sft_2_wo_mask.py ADDED
@@ -0,0 +1,261 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import os
2
+ import copy
3
+ import json
4
+ import logging
5
+ from tqdm import tqdm
6
+ from dataclasses import dataclass, field
7
+ from typing import Optional, Dict, Sequence
8
+
9
+ import torch
10
+ from torch.utils.data import random_split
11
+ from torch.nn.utils.rnn import pad_sequence
12
+ import transformers
13
+ from torch.utils.data import Dataset
14
+ from transformers import Trainer
15
+ import random
16
+ from typing import List, Optional, Tuple, Union
17
+ from transformers import AutoModelForCausalLM, TrainingArguments
18
+ from datasets import load_dataset
19
+ from transformers import DataCollatorForSeq2Seq
20
+ import shutil
21
+
22
+ # from liger_kernel.transformers import AutoLigerKernelForCausalLM
23
+
24
+
25
+ import matplotlib.pyplot as plt
26
+ import numpy as np
27
+
28
+
29
+ @dataclass
30
+ class ModelArguments:
31
+ model_name_or_path: Optional[str] = field(default="facebook/opt-125m")
32
+ # flash_attention: Optional[bool] = field(default=False)
33
+ tokenizer_name_or_path: Optional[str] = field(default=None)
34
+
35
+
36
+ @dataclass
37
+ class DataArguments:
38
+ data_path: str = field(
39
+ default=None, metadata={"help": "Path to the training data."}
40
+ )
41
+ prompt_type: Optional[str] = field(default="instruction")
42
+ dailog_augmentation: Optional[bool] = field(default=False)
43
+
44
+
45
+ @dataclass
46
+ class TrainingArguments(transformers.TrainingArguments):
47
+ cache_dir: Optional[str] = field(default=None)
48
+ optim: str = field(default="adamw_torch")
49
+ model_max_length: int = field(
50
+ default=512,
51
+ metadata={
52
+ "help": "Maximum sequence length. Sequences will be right padded (and possibly truncated)."
53
+ },
54
+ )
55
+
56
+
57
+ IGNORE_INDEX = -100
58
+ MAX_LENGTH = 2000
59
+
60
+ def process(sample, tokenizer):
61
+ # build inputs with format `<bos> X Y <eos>` and labels with format `<ignore> ... <ignore> Y <eos>`
62
+ # for multiturn examples, we only mask the prompt part in each prompt-response pair.
63
+ source = sample["input"]
64
+ # print(source)
65
+ # print(tokenizer.bos_token)
66
+ # print(source == None)
67
+ # if tokenizer.bos_token not in source:
68
+ # source = tokenizer.apply_chat_template(
69
+ # [
70
+ # {'role': 'user', 'content': source}
71
+ # ],
72
+ # tokenize=False, add_generation_prompt=True
73
+ # )
74
+ # # print(source)
75
+ source = tokenizer.apply_chat_template(
76
+ [
77
+ {'role': 'user', 'content': source}
78
+ ],
79
+ tokenize=False, add_generation_prompt=True
80
+ )
81
+ # print(source)
82
+ source = tokenizer(source, add_special_tokens=False)["input_ids"]
83
+ target = [IGNORE_INDEX] * len(source)
84
+ for output in sample["output"]:
85
+ for k, v in output.items():
86
+ if v is None:
87
+ continue
88
+ v_tokens = tokenizer(v, add_special_tokens=False)["input_ids"]
89
+ if k in ["gen"]:
90
+ source += v_tokens
91
+ target += v_tokens
92
+ elif k in ["doc_gen"]:
93
+ source += v_tokens
94
+ target += [IGNORE_INDEX] * len(v_tokens)
95
+ input_ids = source
96
+ labels = target
97
+
98
+ input_ids.append(tokenizer.eos_token_id)
99
+ labels.append(tokenizer.eos_token_id)
100
+ # if tokenizer.eos_token_id not in source:
101
+ # input_ids.append(tokenizer.eos_token_id)
102
+ # if tokenizer.eos_token_id not in labels:
103
+ # labels.append(tokenizer.eos_token_id)
104
+
105
+ # if len(input_ids) > MAX_LENGTH: # 删除过长的数据，防止爆显存
106
+ # return None
107
+ result = {
108
+ "input_ids": input_ids,
109
+ "attention_mask": [1] * len(input_ids),
110
+ "labels": labels,
111
+ }
112
+ # print(result)
113
+ return result
114
+
115
+
116
+ def process_math(sample, tokenizer):
117
+ # build inputs with format `<bos> X Y <eos>` and labels with format `<ignore> ... <ignore> Y <eos>`
118
+ # for multiturn examples, we only mask the prompt part in each prompt-response pair.
119
+ source = sample["input"]
120
+ source = tokenizer.apply_chat_template(
121
+ [
122
+ {'role': 'user', 'content': source}
123
+ ],
124
+ tokenize=False, add_generation_prompt=True
125
+ )
126
+
127
+ source = tokenizer(source, add_special_tokens=False)["input_ids"]
128
+ target = [IGNORE_INDEX] * len(source)
129
+
130
+ output = sample["output_text"]
131
+ output = tokenizer(output, add_special_tokens=False)["input_ids"]
132
+
133
+ source += output
134
+ target += output
135
+
136
+ input_ids = source
137
+ labels = target
138
+
139
+ input_ids.append(tokenizer.eos_token_id)
140
+ labels.append(tokenizer.eos_token_id)
141
+ # if tokenizer.eos_token_id not in source:
142
+ # input_ids.append(tokenizer.eos_token_id)
143
+ # if tokenizer.eos_token_id not in labels:
144
+ # labels.append(tokenizer.eos_token_id)
145
+
146
+ # if len(input_ids) > MAX_LENGTH: # 删除过长的数据，防止爆显存
147
+ # return None
148
+ result = {
149
+ "input_ids": input_ids,
150
+ "attention_mask": [1] * len(input_ids),
151
+ "labels": labels,
152
+ }
153
+ # print(result)
154
+ return result
155
+
156
+
157
+
158
+ def print_function(example, tokenizer):
159
+ print("input_ids:\n{}".format(example["input_ids"]))
160
+ print("inputs:\n{}".format(tokenizer.decode(example["input_ids"], skip_special_tokens=False)))
161
+ print("label_ids:\n{}".format(example["labels"]))
162
+ print("labels:\n{}".format(
163
+ tokenizer.decode(list(filter(lambda x: x != IGNORE_INDEX, example["labels"])), skip_special_tokens=False)
164
+ ))
165
+
166
+
167
+ def get_dataset(file_path, tokenizer, math_dataset=False):
168
+ dataset = load_dataset('json', data_files=file_path)
169
+ train_dataset = dataset["train"]
170
+ file_name = os.path.basename(file_path)
171
+ dataset_name = os.path.splitext(file_name)[0]
172
+ # print(f"dataset_name: {dataset_name}")
173
+ # if os.path.exists(f"input/real_cache/{dataset_name}/"):
174
+ # shutil.rmtree(f"input/real_cache/{dataset_name}/")
175
+ # os.makedirs(f"input/real_cache/{dataset_name}/", exist_ok=True)
176
+ # tokenized_dataset = train_dataset.map(process, fn_kwargs={'tokenizer': tokenizer}, num_proc=1, cache_file_name=f"input/real_cache/{dataset_name}/cache1.arrow")
177
+
178
+ if math_dataset:
179
+ print("math dataste")
180
+ tokenized_dataset = train_dataset.map(process_math, fn_kwargs={'tokenizer': tokenizer}, num_proc=1, load_from_cache_file=False)
181
+ else:
182
+ tokenized_dataset = train_dataset.map(process, fn_kwargs={'tokenizer': tokenizer}, num_proc=1, load_from_cache_file=False)
183
+ print_function(next(iter(tokenized_dataset)), tokenizer)
184
+ print(f"len of dataset before filter: {len(tokenized_dataset)}")
185
+
186
+ print(f"type of tokenized_dataset: {type(tokenized_dataset)}")
187
+
188
+ filtered_dataset = []
189
+ for item in tokenized_dataset:
190
+ # if len(item["input_ids"]) <= 10000:
191
+ # print(item)
192
+ print(f"type of item: {type(item)}")
193
+ filtered_dataset.append(item)
194
+ print(f"len of dataset after filter: {len(filtered_dataset)}")
195
+ return filtered_dataset
196
+
197
+
198
+ def train():
199
+ parser = transformers.HfArgumentParser(
200
+ (ModelArguments, DataArguments, TrainingArguments)
201
+ )
202
+ model_args, data_args, training_args = parser.parse_args_into_dataclasses()
203
+
204
+ print("==========Model Args=========")
205
+ print(model_args)
206
+ print("==========Data Args=========")
207
+ print(data_args)
208
+ print("==========Training Args=========")
209
+ print(training_args)
210
+
211
+ if training_args.gradient_checkpointing:
212
+ use_cache = False # use_cache与gradient_checkpointing不能同时设置为true
213
+ model = AutoModelForCausalLM.from_pretrained(
214
+ model_args.model_name_or_path,
215
+ _attn_implementation="flash_attention_2",
216
+ use_cache=use_cache,
217
+ # save_only_model=True
218
+ ).float()
219
+ # model = AutoLigerKernelForCausalLM.from_pretrained(
220
+ # model_args.model_name_or_path,
221
+ # _attn_implementation="flash_attention_2",
222
+ # use_cache=use_cache,
223
+ # # save_only_model=True
224
+ # ).float()
225
+ if model_args.tokenizer_name_or_path is None:
226
+ model_args.tokenizer_name_or_path = model_args.model_name_or_path
227
+ tokenizer = transformers.AutoTokenizer.from_pretrained(
228
+ model_args.tokenizer_name_or_path, model_max_length=training_args.model_max_length
229
+ )
230
+
231
+ if tokenizer.pad_token is None:
232
+ tokenizer.pad_token = tokenizer.eos_token
233
+
234
+ dataset_qa = get_dataset(data_args.data_path, tokenizer, True)
235
+ print(f"qa dataset: {len(dataset_qa)}")
236
+ # dataset_math = get_dataset("/opt/aps/workdir/sunshuang/deep_search/search_o1/sft_data/math_qwq_4524_selected_add_prompt_871.json", tokenizer, True)
237
+ # print(f"math dataset: {len(dataset_math)}")
238
+ dataset = dataset_qa
239
+ # dataset = get_dataset(data_args.data_path, tokenizer)
240
+ print(f"fianl dataset: {len(dataset)}")
241
+
242
+ data_collator = DataCollatorForSeq2Seq(
243
+ tokenizer=tokenizer,
244
+ label_pad_token_id=-100,
245
+ )
246
+
247
+ trainer = Trainer(
248
+ model=model,
249
+ args=training_args,
250
+ tokenizer=tokenizer,
251
+ data_collator=data_collator,
252
+ train_dataset=dataset,
253
+ )
254
+ trainer.train(resume_from_checkpoint=training_args.resume_from_checkpoint)
255
+ trainer.save_model(training_args.output_dir)
256
+ trainer.save_state()
257
+
258
+
259
+ if __name__ == "__main__":
260
+ torch.manual_seed(42)
261
+ train()
deep_search/sft/test.ipynb ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "cells": [
3
+ {
4
+ "cell_type": "code",
5
+ "execution_count": null,
6
+ "metadata": {},
7
+ "outputs": [],
8
+ "source": [
9
+ "import os\n",
10
+ "from transformers import AutoModelForCausalLM, AutoTokenizer\n",
11
+ "# os.environ[\"CUDA_VISIBLE_DEVICES\"] = '6,7'\n",
12
+ "model_path = \"/capacity/userdata/models/DeepSeek-R1-Distill-Qwen-32B\"\n",
13
+ "\n",
14
+ "# model = AutoModelForCausalLM.from_pretrained(model_path, device_map='auto')\n",
15
+ "tokenizer = AutoTokenizer.from_pretrained(model_path)\n",
16
+ "\n",
17
+ "inputs = tokenizer.apply_chat_template(\n",
18
+ " [\n",
19
+ " {'role': 'user', 'content': \"I hate you\"}\n",
20
+ " ],\n",
21
+ " tokenize=False, add_generation_prompt=True\n",
22
+ ")\n",
23
+ "print(inputs)"
24
+ ]
25
+ },
26
+ {
27
+ "cell_type": "code",
28
+ "execution_count": null,
29
+ "metadata": {},
30
+ "outputs": [],
31
+ "source": [
32
+ "import json\n",
33
+ "\n",
34
+ "\n",
35
+ "\n",
36
+ "def load_json(file_path):\n",
37
+ " with open(file_path, \"r\", encoding=\"utf-8\") as file:\n",
38
+ " data = json.load(file)\n",
39
+ " print(f\"Loaded {len(data)} items from {file_path}\")\n",
40
+ " return data\n",
41
+ "\n",
42
+ "def save_json(data, file_path):\n",
43
+ " with open(file_path, \"w\", encoding=\"utf-8\") as file:\n",
44
+ " json.dump(data, file, ensure_ascii=False, indent=4)\n",
45
+ " print(f\"Saved {len(data)} items to {file_path}\")\n",
46
+ "\n",
47
+ "data = load_json(\"/share/project/sunshuang/deep_search/search_o1/sft_data/new_instruction_2k_sft.json\")"
48
+ ]
49
+ }
50
+ ],
51
+ "metadata": {
52
+ "language_info": {
53
+ "name": "python"
54
+ }
55
+ },
56
+ "nbformat": 4,
57
+ "nbformat_minor": 2
58
+ }
deep_search/sft/test_len.sh ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ # qwen 7b 用自己的tokenizer
3
+ # /opt/aps/workdir/miniforge3/envs/train/bin/deepspeed \
4
+ # --master_port=9944 \
5
+ # --include localhost:6,7 \
6
+ # sft.py \
7
+ # --deepspeed ds_zero3.json \
8
+ # --model_name_or_path /opt/aps/workdir/home/.cache/modelscope/hub/models/qwen/Qwen2.5-7B-Instruct \
9
+ # --tokenizer_name_or_path /opt/aps/workdir/home/.cache/modelscope/hub/models/qwen/Qwen2.5-7B-Instruct \
10
+ # --do_train \
11
+ # --save_safetensors true \
12
+ # --data_path /opt/aps/workdir/input/data/search-o1/hopotqa_1217.json \
13
+ # --lr_scheduler_type cosine \
14
+ # --output_dir output/sft_use_original_tokenizer/qwen_7b_original_tokenizer_inst_data_1217_1 \
15
+ # --overwrite_output_dir \
16
+ # --warmup_ratio 0.03 \
17
+ # --gradient_checkpointing true \
18
+ # --per_device_train_batch_size 1 \
19
+ # --gradient_accumulation_steps 4 \
20
+ # --logging_steps 1 \
21
+ # --learning_rate 2e-5 \
22
+ # --num_train_epochs 1 \
23
+ # --model_max_length 8192 \
24
+ # --save_total_limit 16 \
25
+ # --bf16 || exit 1
deep_search/sft/test_two_model_qwq.sh ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ model_save_path=$1
2
+ model_name=$2
3
+ dataset=$3
4
+ # cd /share/project/sunshuang/deep_search/search_o1
5
+
6
+ # CUDA_VISIBLE_DEVICES=0,1 nohup \
7
+ # /opt/aps/workdir/miniforge3/envs/search_o1/bin/vllm serve /capacity/userdata/models/Qwen2.5-32B-Instruct \
8
+ # --tensor-parallel-size=2 \
9
+ # --gpu-memory-utilization 0.95 \
10
+ # --port 8001 \
11
+ # > vllm-Qwen32B.log 2>&1 &
12
+ # tail -f vllm-Qwen32B.log
13
+ # exit 1
14
+
15
+ # sleep 300 # 等待5min
16
+
17
+
18
+ cd /opt/aps/workdir/sunshuang/deep_search/sft
19
+
20
+ echo "start eval ${model_save_path}"
21
+
22
+ # /share/project/miniconda/envs/search_o1/bin/python -u wait_eval_1.py ${model_save_path} > 3-5-2013qa_500math_doc_by_qwen.log 2>&1
23
+
24
+
25
+
26
+ sleep 300 # 等待5min
27
+
28
+ echo "vllm serve qwq"
29
+
30
+ export CUDA_VISIBLE_DEVICES=0,1
31
+ nohup vllm serve /capacity/userdata/models/QwQ-32B \
32
+ --tensor-parallel-size=2 \
33
+ --gpu-memory-utilization 0.95 \
34
+ --port 8003 > /opt/aps/workdir/sunshuang/deep_search/search_o1/sft_logs/vllm_serve_qwq.log 2>&1 &
35
+
36
+ echo "finish vllm serve qwq"
37
+
38
+ sleep 300
39
+
40
+ nohup /opt/aps/workdir/search_o1/bin/python3 -u /opt/aps/workdir/sunshuang/deep_search/sft/wait_eval_use_one_model_for_ckpt_two_model_qwq.py ${model_save_path} > 5-5_${dataset}_${model_name}.log 2>&1 &
41
+
42
+
43
+ nohup /opt/aps/workdir/search_o1/bin/python3 -u /opt/aps/workdir/sunshuang/deep_search/sft/wait_eval_use_one_model_for_ckpt_two_model_qwq.py /capacity/userdata/ss/sft_search/JOB:15751#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_ablation_subquery_1073_random_sample_871_ablation > 5-6_1.log 2>&1 &
44
+
45
+
46
+ # nohup /opt/aps/workdir/search_o1/bin/python3 -u /opt/aps/workdir/sunshuang/deep_search/sft/wait_eval_use_one_model_for_ckpt_two_model.py /capacity/userdata/ss/sft_search/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871 > 4-24_no_error_data_871_doc_by_itself_Qwen2.5-7B-Instruct_add_math871_2.log 2>&1 &
47
+ # nohup /opt/aps/workdir/search_o1/bin/python3 -u /opt/aps/workdir/sunshuang/deep_search/sft/wait_eval_use_one_model_for_ckpt_two_model.py /capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871 > 4-24_math_qwq_4524_selected_add_prompt_871_doc_by_itself_qwen7b_sft_871_checkpoint-78_add_math871_after_search_2.log 2>&1 &
48
+ # nohup /share/project/miniconda/envs/search_o1/bin/python -u wait_eval_use_one_model_for_ckpt.py /share/project/sunshuang/deep_search/output/JOB:2934#LR:1e-5#BASE:Qwen2.5-32B-Instruct#TOKEN:Qwen2.5-32B-Instruct#BSZ:2#ACC:4_strict_selected_1526_sft > 3-25-strict_selected_1526_sft_doc_by_itself_Qwen2.5-32B-Instruct_eval.log 2>&1 &
49
+ # /share/project/sunshuang/deep_search/output/JOB:13192#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_strict_selected_1526_sft
50
+ # /share/project/sunshuang/deep_search/output/JOB:2934#LR:1e-5#BASE:Qwen2.5-32B-Instruct#TOKEN:Qwen2.5-32B-Instruct#BSZ:2#ACC:4_strict_selected_1526_sft
51
+ # # nohup /share/project/miniconda/envs/search_o1/bin/python -u wait_eval_use_one_model_for_ckpt_check.py /share/project/sunshuang/deep_search/output/JOB:11783#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_selected_data_1174_sft > 3-26_selected_data_1174_sft_doc_by_itself_QwQ-32B_2.log 2>&1 &
52
+ # nohup /opt/aps/workdir/search_o1/bin/python3 -u wait_eval_use_one_model_for_ckpt.py /capacity/userdata/ss/sft_search/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469 > 3-28_merged_1174_zh_296_sft_1469_doc_by_itself_QwQ-32B.log 2>&1 &
53
+ # nohup /opt/aps/workdir/search_o1/bin/python3 -u wait_eval_use_one_model_for_ckpt.py /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829 > 3-28_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829_doc_by_itself_QwQ-32B_3.log 2>&1 &
54
+ # nohup /opt/aps/workdir/search_o1/bin/python3 -u /opt/aps/workdir/sunshuang/deep_search/sft/wait_eval_use_one_model_for_ckpt_worker1.py /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097 > 4-1_1.1k_cleaned_data_1097_doc_by_itself_QwQ-32B_3.log 2>&1 &
55
+ # nohup /opt/aps/workdir/search_o1/bin/python3 -u /opt/aps/workdir/sunshuang/deep_search/sft/wait_eval_use_one_model_for_ckpt.py /capacity/userdata/ss/sft_search/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871 > 4-4_no_error_data_871_doc_by_itself_DeepSeek-R1-Distill-Qwen-32.log 2>&1 &
56
+
57
+ # nohup /share/project/miniconda/envs/search_o1/bin/python -u wait_eval_use_one_model_for_ckpt.py /share/project/sunshuang/deep_search/output/JOB:27033#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_syn_long_359_sft_1533 > 3-27_merged_syn_long_359_sft_1533_doc_by_itself_QwQ-32B_1.log 2>&1 &
58
+
59
+ # nohup /share/project/miniconda/envs/search_o1/bin/python -u wait_eval_use_one_model_for_ckpt.py /share/project/sunshuang/deep_search/output/JOB:2401#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_strict_selected_1526_sft_format_ans > 3-26_strict_selected_1526_sft_format_ans_doc_by_itself_QwQ-32B_1.log 2>&1 &
60
+
61
+
62
+ # nohup /share/project/miniconda/envs/search_o1/bin/python -u wait_eval_use_one_model_for_ckpt.py /share/project/sunshuang/deep_search/output/JOB:12665#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_new_instruction_1161_1301_sft_remove_exp_2462 > 3-20-merged_2462_remove_exp_doc_by_itself_qwq.log 2>&1 &
63
+
64
+ # nohup /share/project/miniconda/envs/search_o1/bin/python -u wait_eval_use_one_model_for_ckpt.py /share/project/sunshuang/deep_search/output/JOB:487#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32B#TOKEN:DeepSeek-R1-Distill-Qwen-32B#BSZ:2#ACC:4 > 3-12-1200qa_doc_by_itself_r1_32b.log 2>&1 &
deep_search/sft/test_two_model_qwq_1.sh ADDED
@@ -0,0 +1,69 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ model_save_path=$1
2
+ model_name=$2
3
+ dataset=$3
4
+ # cd /share/project/sunshuang/deep_search/search_o1
5
+
6
+ # CUDA_VISIBLE_DEVICES=0,1 nohup \
7
+ # /opt/aps/workdir/miniforge3/envs/search_o1/bin/vllm serve /capacity/userdata/models/Qwen2.5-32B-Instruct \
8
+ # --tensor-parallel-size=2 \
9
+ # --gpu-memory-utilization 0.95 \
10
+ # --port 8001 \
11
+ # > vllm-Qwen32B.log 2>&1 &
12
+ # tail -f vllm-Qwen32B.log
13
+ # exit 1
14
+
15
+ # sleep 300 # 等待5min
16
+
17
+
18
+ cd /opt/aps/workdir/sunshuang/deep_search/sft
19
+
20
+ echo "start eval ${model_save_path}"
21
+
22
+ # /share/project/miniconda/envs/search_o1/bin/python -u wait_eval_1.py ${model_save_path} > 3-5-2013qa_500math_doc_by_qwen.log 2>&1
23
+
24
+
25
+
26
+ sleep 300 # 等待5min
27
+
28
+ # echo "vllm serve qwq".
29
+
30
+ # export CUDA_VISIBLE_DEVICES=0,1
31
+ # nohup vllm serve /capacity/userdata/models/QwQ-32B \
32
+ # --tensor-parallel-size=2 \
33
+ # --gpu-memory-utilization 0.95 \
34
+ # --port 8003 > /opt/aps/workdir/sunshuang/deep_search/search_o1/sft_logs/vllm_serve_qwq_1.log 2>&1 &
35
+
36
+ # # echo "finish vllm serve qwq"
37
+
38
+ # sleep 300
39
+
40
+
41
+ nohup /opt/aps/workdir/search_o1/bin/python3 -u /opt/aps/workdir/sunshuang/deep_search/sft/wait_eval_use_one_model_for_ckpt_two_model_qwq.py ${model_save_path} > 5-5_${dataset}_${model_name}.log 2>&1 &
42
+
43
+
44
+
45
+ # nohup /opt/aps/workdir/search_o1/bin/python3 -u /opt/aps/workdir/sunshuang/deep_search/sft/wait_eval_use_one_model_for_ckpt_two_model_qwq_1.py ${model_save_path} > 4-25_${dataset}_doc_by_itself_${model_name}_1.log 2>&1 &
46
+
47
+ # nohup /opt/aps/workdir/search_o1/bin/python3 -u /opt/aps/workdir/sunshuang/deep_search/sft/wait_eval_use_one_model_for_ckpt_two_model_qwq_1.py /capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math > 4-25_no_error_data_871_doc_by_itself_QwQ-32B_1.log 2>&1 &
48
+
49
+ # nohup /opt/aps/workdir/search_o1/bin/python3 -u /opt/aps/workdir/sunshuang/deep_search/sft/wait_eval_use_one_model_for_ckpt_two_model_qwq.py /capacity/userdata/ss/sft_search/JOB:16104#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:1#ACC:8_no_error_data_871_mixed_math > 4-25_no_error_data_871_doc_by_itself_QwQ-32B.log 2>&1 &
50
+
51
+ # nohup /opt/aps/workdir/search_o1/bin/python3 -u /opt/aps/workdir/sunshuang/deep_search/sft/wait_eval_use_one_model_for_ckpt_two_model.py /capacity/userdata/ss/sft_search/JOB:30936#LR:1e-5#BASE:Qwen2.5-7B-Instruct#TOKEN:Qwen2.5-7B-Instruct#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871 > 4-24_no_error_data_871_doc_by_itself_Qwen2.5-7B-Instruct_add_math871_2.log 2>&1 &
52
+ # nohup /opt/aps/workdir/search_o1/bin/python3 -u /opt/aps/workdir/sunshuang/deep_search/sft/wait_eval_use_one_model_for_ckpt_two_model.py /capacity/userdata/ss/sft_search/JOB:16641#LR:1e-5#BASE:qwen7b_sft_871_checkpoint-78#TOKEN:qwen7b_sft_871_checkpoint-78#BSZ:2#ACC:4_math_qwq_4524_selected_add_prompt_871 > 4-24_math_qwq_4524_selected_add_prompt_871_doc_by_itself_qwen7b_sft_871_checkpoint-78_add_math871_after_search_2.log 2>&1 &
53
+ # nohup /share/project/miniconda/envs/search_o1/bin/python -u wait_eval_use_one_model_for_ckpt.py /share/project/sunshuang/deep_search/output/JOB:2934#LR:1e-5#BASE:Qwen2.5-32B-Instruct#TOKEN:Qwen2.5-32B-Instruct#BSZ:2#ACC:4_strict_selected_1526_sft > 3-25-strict_selected_1526_sft_doc_by_itself_Qwen2.5-32B-Instruct_eval.log 2>&1 &
54
+ # /share/project/sunshuang/deep_search/output/JOB:13192#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_strict_selected_1526_sft
55
+ # /share/project/sunshuang/deep_search/output/JOB:2934#LR:1e-5#BASE:Qwen2.5-32B-Instruct#TOKEN:Qwen2.5-32B-Instruct#BSZ:2#ACC:4_strict_selected_1526_sft
56
+ # # nohup /share/project/miniconda/envs/search_o1/bin/python -u wait_eval_use_one_model_for_ckpt_check.py /share/project/sunshuang/deep_search/output/JOB:11783#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_selected_data_1174_sft > 3-26_selected_data_1174_sft_doc_by_itself_QwQ-32B_2.log 2>&1 &
57
+ # nohup /opt/aps/workdir/search_o1/bin/python3 -u wait_eval_use_one_model_for_ckpt.py /capacity/userdata/ss/sft_search/JOB:30386#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_zh_296_sft_1469 > 3-28_merged_1174_zh_296_sft_1469_doc_by_itself_QwQ-32B.log 2>&1 &
58
+ # nohup /opt/aps/workdir/search_o1/bin/python3 -u wait_eval_use_one_model_for_ckpt.py /capacity/userdata/ss/sft_search/JOB:9446#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829 > 3-28_merged_1174_syn_long_359_sft_1533_zh_296_sft_1829_doc_by_itself_QwQ-32B_3.log 2>&1 &
59
+ # nohup /opt/aps/workdir/search_o1/bin/python3 -u /opt/aps/workdir/sunshuang/deep_search/sft/wait_eval_use_one_model_for_ckpt_worker1.py /capacity/userdata/ss/sft_search/JOB:17755#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_1.1k_cleaned_data_1097 > 4-1_1.1k_cleaned_data_1097_doc_by_itself_QwQ-32B_3.log 2>&1 &
60
+ # nohup /opt/aps/workdir/search_o1/bin/python3 -u /opt/aps/workdir/sunshuang/deep_search/sft/wait_eval_use_one_model_for_ckpt.py /capacity/userdata/ss/sft_search/JOB:17315#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32#TOKEN:DeepSeek-R1-Distill-Qwen-32#BSZ:2#ACC:4_no_error_data_871 > 4-4_no_error_data_871_doc_by_itself_DeepSeek-R1-Distill-Qwen-32.log 2>&1 &
61
+
62
+ # nohup /share/project/miniconda/envs/search_o1/bin/python -u wait_eval_use_one_model_for_ckpt.py /share/project/sunshuang/deep_search/output/JOB:27033#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_syn_long_359_sft_1533 > 3-27_merged_syn_long_359_sft_1533_doc_by_itself_QwQ-32B_1.log 2>&1 &
63
+
64
+ # nohup /share/project/miniconda/envs/search_o1/bin/python -u wait_eval_use_one_model_for_ckpt.py /share/project/sunshuang/deep_search/output/JOB:2401#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_strict_selected_1526_sft_format_ans > 3-26_strict_selected_1526_sft_format_ans_doc_by_itself_QwQ-32B_1.log 2>&1 &
65
+
66
+
67
+ # nohup /share/project/miniconda/envs/search_o1/bin/python -u wait_eval_use_one_model_for_ckpt.py /share/project/sunshuang/deep_search/output/JOB:12665#LR:1e-5#BASE:QwQ-32B#TOKEN:QwQ-32B#BSZ:2#ACC:4_merged_new_instruction_1161_1301_sft_remove_exp_2462 > 3-20-merged_2462_remove_exp_doc_by_itself_qwq.log 2>&1 &
68
+
69
+ # nohup /share/project/miniconda/envs/search_o1/bin/python -u wait_eval_use_one_model_for_ckpt.py /share/project/sunshuang/deep_search/output/JOB:487#LR:1e-5#BASE:DeepSeek-R1-Distill-Qwen-32B#TOKEN:DeepSeek-R1-Distill-Qwen-32B#BSZ:2#ACC:4 > 3-12-1200qa_doc_by_itself_r1_32b.log 2>&1 &
deep_search/sft/train-inst.py ADDED
@@ -0,0 +1,311 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import copy
2
+ import json
3
+ import logging
4
+ from dataclasses import dataclass, field
5
+ from typing import Optional, Dict, Sequence
6
+
7
+ import torch
8
+ import torch.nn as nn
9
+ from torch.nn.utils.rnn import pad_sequence
10
+ import transformers
11
+ from torch.utils.data import Dataset
12
+ from transformers import Trainer
13
+ import random
14
+ from typing import List, Optional, Tuple, Union
15
+ from transformers.models.llama.modeling_llama import (
16
+ LlamaForCausalLM,
17
+ LlamaAttention,
18
+ apply_rotary_pos_emb,
19
+ )
20
+
21
+
22
+ IGNORE_INDEX = -100
23
+ DEFAULT_PAD_TOKEN = "[PAD]"
24
+ DEFAULT_EOS_TOKEN = "</s>"
25
+ DEFAULT_BOS_TOKEN = "<s>"
26
+ DEFAULT_UNK_TOKEN = "[UNK]"
27
+
28
+
29
+
30
+ def apply_template_self(example, tokenizer):
31
+ system_prompt='''You are a helpful assistant. Given a question, you should answer it by first thinking about the reasoning process in the mind and then providing the final answer. The output format of reasoning process and final answer are enclosed within <think> </think> and <answer> </answer> tags, respectively, i.e., "<think> reasoning process here </think>\n\n<answer> final answer here </answer>". During the thinking process, you can perform searching for uncertain knowledge if necessary with the format of "<|begin_of_query|> search query (only keywords) here <|end_of_query|>". Then, the system will provide the Assistant with helpful information with the format of "<|begin_of_documents|> ...search results... <|end_of_documents|>"'''
32
+
33
+ prompt = tokenizer.apply_chat_template(
34
+ [
35
+ {"role": "system", "content": system_prompt},
36
+ {"role": "user", "content": example},
37
+ ],
38
+ tokenize=False,
39
+ add_generation_prompt=True,
40
+ )
41
+ # print(prompt)
42
+ return prompt
43
+
44
+
45
+
46
+
47
+ def apply_template(example, tokenizer):
48
+ system_prompt="You are skilled at solving problems through step-by-step reasoning, and you can choose different actions to address the problems."
49
+ user_prompt='''Use the following five actions to solve the problem step by step:
50
+ <Sub-question Decompose>: Plan the next sub-question that needs to be solved.
51
+ <Sub-question Answer>: Directly answer or utilize the retrieved documents to answer a sub-question.
52
+ <Search>: Utilize external the external Search Engine to retrieve knowledge. **When calling the search engine, you need to encapsulate the query keywords within <|begin_of_query|><|end_of_query|> tags**.
53
+ <Reflect>: Reflect on past actions and their outcomes.
54
+ <Final Answer>: Obtain the final answer to the original complex question, and **output the final answer in \\boxed{{}}**.
55
+
56
+ ## Question
57
+ {question}'''
58
+
59
+ prompt = tokenizer.apply_chat_template(
60
+ [
61
+ {"role": "system", "content": system_prompt},
62
+ {"role": "user", "content": user_prompt.replace("{question}",example)},
63
+ ],
64
+ tokenize=False,
65
+ add_generation_prompt=True,
66
+ )
67
+ prompt = prompt+"\n## Solution"
68
+ # print(prompt)
69
+ # kill
70
+ return prompt
71
+
72
+
73
+
74
+ @dataclass
75
+ class ModelArguments:
76
+ model_name_or_path: Optional[str] = field(default="facebook/opt-125m")
77
+ flash_attention: Optional[bool] = field(default=False)
78
+
79
+
80
+ @dataclass
81
+ class DataArguments:
82
+ data_path: str = field(
83
+ default=None, metadata={"help": "Path to the training data."}
84
+ )
85
+ prompt_type: Optional[str] = field(default="instruction")
86
+ dailog_augmentation: Optional[bool] = field(default=False)
87
+
88
+
89
+ @dataclass
90
+ class TrainingArguments(transformers.TrainingArguments):
91
+ cache_dir: Optional[str] = field(default=None)
92
+ optim: str = field(default="adamw_torch")
93
+ model_max_length: int = field(
94
+ default=512,
95
+ metadata={
96
+ "help": "Maximum sequence length. Sequences will be right padded (and possibly truncated)."
97
+ },
98
+ )
99
+
100
+
101
+ def safe_save_model_for_hf_trainer(trainer: transformers.Trainer, output_dir: str):
102
+ """Collects the state dict and dump to disk."""
103
+ state_dict = trainer.model.state_dict()
104
+ if trainer.args.should_save:
105
+ cpu_state_dict = {key: value.cpu() for key, value in state_dict.items()}
106
+ del state_dict
107
+ trainer._save(output_dir, state_dict=cpu_state_dict) # noqa
108
+
109
+
110
+ def smart_tokenizer_and_embedding_resize(
111
+ special_tokens_dict: Dict,
112
+ tokenizer: transformers.PreTrainedTokenizer,
113
+ model: transformers.PreTrainedModel,
114
+ ):
115
+ """Resize tokenizer and embedding.
116
+
117
+ Note: This is the unoptimized version that may make your embedding size not be divisible by 64.
118
+ """
119
+ num_new_tokens = tokenizer.add_special_tokens(special_tokens_dict)
120
+ model.resize_token_embeddings(len(tokenizer))
121
+
122
+ if num_new_tokens > 0:
123
+ input_embeddings = model.get_input_embeddings().weight.data
124
+ output_embeddings = model.get_output_embeddings().weight.data
125
+
126
+ input_embeddings_avg = input_embeddings[:-num_new_tokens].mean(
127
+ dim=0, keepdim=True
128
+ )
129
+ output_embeddings_avg = output_embeddings[:-num_new_tokens].mean(
130
+ dim=0, keepdim=True
131
+ )
132
+
133
+ input_embeddings[-num_new_tokens:] = input_embeddings_avg
134
+ output_embeddings[-num_new_tokens:] = output_embeddings_avg
135
+
136
+
137
+ class SupervisedDataset(Dataset):
138
+ """Dataset for supervised fine-tuning."""
139
+
140
+ def __init__(
141
+ self,
142
+ data_path: str,
143
+ prompt_type: str,
144
+ tokenizer: transformers.PreTrainedTokenizer,
145
+ ):
146
+ super(SupervisedDataset, self).__init__()
147
+ logging.warning("Loading data...")
148
+ # prompt_simple_inference = PROMPT_DICT["alpaca_format"]
149
+ self.sources, self.targets = [], []
150
+ for path in data_path.split(","):
151
+ with open(path, "r") as f:
152
+ for i, line in enumerate(f.readlines()):
153
+ try:
154
+ c = json.loads(line)
155
+ except:
156
+ print(path)
157
+ print(line)
158
+ raise ValueError
159
+
160
+ if "question" in c:
161
+ input_text = c["question"]
162
+ source = apply_template(input_text, tokenizer)
163
+ else:
164
+ source = c["prompt"]
165
+ self.sources.append(source.strip())
166
+
167
+ if "solution" in c:
168
+ output_text = c["solution"]
169
+ else:
170
+ output_text = c["chosen"]
171
+ self.targets.append(source + output_text + tokenizer.eos_token)
172
+ print("Train Dataset Length: ",len(self.sources))
173
+
174
+ def __len__(self):
175
+ return len(self.sources)
176
+
177
+ def __getitem__(self, i) -> Dict[str, torch.Tensor]:
178
+ return dict(input_ids=self.sources[i], labels=self.targets[i])
179
+
180
+
181
+ @dataclass
182
+ class DataCollatorForSupervisedDataset(object):
183
+ """Collate examples for supervised fine-tuning."""
184
+
185
+ data_args: DataArguments
186
+ tokenizer: transformers.PreTrainedTokenizer
187
+
188
+ def __call__(self, instances: Sequence[Dict]) -> Dict[str, torch.Tensor]:
189
+ # print(instances)
190
+ # print("======="*30)
191
+
192
+ inputs = self.tokenizer(
193
+ text=[instance["labels"] for instance in instances],
194
+ text_target=[instance["input_ids"] for instance in instances],
195
+ return_tensors="pt",
196
+ padding="longest",
197
+ max_length=self.tokenizer.model_max_length,
198
+ truncation=True,
199
+ return_attention_mask=True,
200
+ )
201
+ labels = copy.deepcopy(inputs["input_ids"])
202
+ labels[labels == self.tokenizer.pad_token_id] = IGNORE_INDEX
203
+ labels[torch.where(inputs["labels"] != self.tokenizer.pad_token_id)] = (
204
+ IGNORE_INDEX
205
+ )
206
+
207
+ inputs["labels"] = labels
208
+
209
+ for k in range(len(labels)):
210
+ label = torch.clone(inputs['labels'][k])
211
+ input_id = torch.clone(inputs['input_ids'][k])
212
+ attention_mask = torch.clone(inputs['attention_mask'][k])
213
+ # print("Label:",label.tolist())
214
+ # print("Input-ID:",input_id.tolist())
215
+
216
+ start_tokens = torch.tensor([27, 91, 7265, 3575, 75927, 91, 397]) # <|begin_of_documents|>\n
217
+ end_tokens = torch.tensor([408, 3575, 75927, 91, 1339])
218
+
219
+ is_in_masking = False
220
+ mask_start_idx = -1
221
+ start_count = 0
222
+ end_count = 0
223
+ for m in range(len(input_id)):
224
+ if label[m] == 0:
225
+ continue
226
+
227
+ if not is_in_masking and m + len(start_tokens) <= len(input_id):
228
+ if torch.equal(input_id[m:m + len(start_tokens)], start_tokens):
229
+ start_count += 1
230
+ is_in_masking = True
231
+ mask_start_idx = m
232
+ label[m:m + len(start_tokens)] = -100
233
+
234
+ if is_in_masking:
235
+ label[m] = -100
236
+ if m + len(end_tokens) <= len(input_id):
237
+ if torch.equal(input_id[m:m + len(end_tokens)], end_tokens):
238
+ end_count += 1 # 增加 end_tokens 计数
239
+ is_in_masking = False # 结束 mask
240
+ label[m:m + len(end_tokens)] = -100
241
+ mask_start_idx = -1
242
+ if start_count != end_count:
243
+ print("Important Error!",start_count,end_count)
244
+ raise ValueError("128017 and 128018 must be in pairs.")
245
+
246
+
247
+ inputs['labels'][k] = label
248
+ inputs['input_ids'][k] = input_id
249
+ inputs['attention_mask'][k] = attention_mask
250
+
251
+ torch.set_printoptions(profile="full")
252
+ # print(inputs)
253
+
254
+ # exit()
255
+
256
+ return inputs
257
+
258
+
259
+ def make_supervised_data_module(
260
+ tokenizer: transformers.PreTrainedTokenizer, data_args
261
+ ) -> Dict:
262
+ """Make dataset and collator for supervised fine-tuning."""
263
+ train_dataset = SupervisedDataset(
264
+ tokenizer=tokenizer,
265
+ prompt_type=data_args.prompt_type,
266
+ data_path=data_args.data_path,
267
+ )
268
+ print(train_dataset)
269
+ data_collator = DataCollatorForSupervisedDataset(
270
+ data_args=data_args, tokenizer=tokenizer
271
+ )
272
+ return dict(
273
+ train_dataset=train_dataset, eval_dataset=None, data_collator=data_collator
274
+ )
275
+
276
+
277
+ def train():
278
+ parser = transformers.HfArgumentParser(
279
+ (ModelArguments, DataArguments, TrainingArguments)
280
+ )
281
+ model_args, data_args, training_args = parser.parse_args_into_dataclasses()
282
+
283
+ model = transformers.AutoModelForCausalLM.from_pretrained(
284
+ model_args.model_name_or_path,
285
+ cache_dir=training_args.cache_dir,
286
+ )
287
+
288
+ tokenizer = transformers.AutoTokenizer.from_pretrained(
289
+ model_args.model_name_or_path,
290
+ cache_dir=training_args.cache_dir,
291
+ model_max_length=training_args.model_max_length,
292
+ padding_side="right",
293
+ )
294
+
295
+ if "llama" in model_args.model_name_or_path.lower():
296
+ tokenizer.pad_token=tokenizer.eos_token
297
+
298
+ data_module = make_supervised_data_module(tokenizer=tokenizer, data_args=data_args)
299
+
300
+ trainer = Trainer(
301
+ model=model, tokenizer=tokenizer, args=training_args, **data_module
302
+ )
303
+ trainer.train(resume_from_checkpoint=training_args.resume_from_checkpoint)
304
+ trainer.save_state()
305
+ trainer.save_model(output_dir=training_args.output_dir)
306
+ # safe_save_model_for_hf_trainer(trainer=trainer, output_dir=training_args.output_dir)
307
+
308
+
309
+ if __name__ == "__main__":
310
+ print("222222")
311
+ train()
deep_search/sft/train-inst.sh ADDED
@@ -0,0 +1,56 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/bin/bash
2
+
3
+ # cd /opt/aps/workdir/songhuatong
4
+ PWD=/opt/aps/workdir/sht-RAG_SFT/resultes
5
+
6
+ # source /opt/aps/workdir/songhuatong/.venv/bin/activate
7
+
8
+
9
+ BS=$1
10
+ ACC=$2
11
+ LR=$3
12
+ WARM_UP_RATIO=$4
13
+ DATA_PATH=$5
14
+
15
+ SCRIPT_PATH=/opt/aps/workdir/sht-RAG_SFT/train-inst.py
16
+ MODEL_PATH=/opt/aps/workdir/model/Qwen2.5-7B-Instruct
17
+ DATA_PREFIX=$(basename $DATA_PATH .jsonl)
18
+ JOB_NAME=qwen-2_5-7B-inst_${DATA_PREFIX}_${BS}_${ACC}_${LR}_${WARM_UP_RATIO}
19
+ SAVE_DIR=${PWD}/model
20
+
21
+ export OMP_NUM_THREADS=24
22
+ export WANDB_MODE=offline
23
+ export CUDA_VISIBLE_DEVICES=4,5,6,7
24
+ echo 1111
25
+ echo ${SCRIPT_PATH}
26
+ torchrun --nproc_per_node 4 \
27
+ --nnodes 1 \
28
+ --node_rank 0 \
29
+ --master_addr lmlabide-6427f7a2-728d-4097-a672-471aaf8c507a-worker-0 \
30
+ --master_port 1111 \
31
+ /opt/aps/workdir/sht-RAG_SFT/train-inst.py \
32
+ --model_name_or_path ${MODEL_PATH} \
33
+ --data_path ${DATA_PATH} \
34
+ --bf16 True \
35
+ --output_dir ${SAVE_DIR}/${JOB_NAME} \
36
+ --num_train_epochs 3 \
37
+ --per_device_train_batch_size ${BS} \
38
+ --per_device_eval_batch_size 1 \
39
+ --gradient_accumulation_steps ${ACC} \
40
+ --evaluation_strategy "no" \
41
+ --save_strategy "epoch" \
42
+ --save_total_limit 1 \
43
+ --learning_rate ${LR} \
44
+ --weight_decay 0. \
45
+ --warmup_ratio ${WARM_UP_RATIO} \
46
+ --lr_scheduler_type "cosine" \
47
+ --logging_steps 1 \
48
+ --model_max_length 8192 \
49
+ --deepspeed /opt/aps/workdir/sht-RAG_SFT/dp_3.jsonl \
50
+ --gradient_checkpointing True \
51
+ --report_to none \
52
+ --tf32 True \
53
+ # &> /home/songhuatong/SFT/logs/${JOB_NAME}.log
54
+
55
+ # 四张卡
56
+ # bash /opt/aps/workdir/sht-RAG_SFT/train-inst.sh 2 4 2e-5 0.1 /opt/aps/workdir/sht-RAG_RL/datasets/rollout/sft_data/mixed_sft.jsonl
deep_search/sft/train_env.yml ADDED
@@ -0,0 +1,145 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ name: ss_train
2
+ channels:
3
+ - https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge
4
+ - conda-forge
5
+ dependencies:
6
+ - _libgcc_mutex=0.1=conda_forge
7
+ - _openmp_mutex=4.5=2_gnu
8
+ - bzip2=1.0.8=h4bc722e_7
9
+ - ca-certificates=2025.1.31=hbcca054_0
10
+ - ld_impl_linux-64=2.43=h712a8e2_2
11
+ - libexpat=2.6.4=h5888daf_0
12
+ - libffi=3.4.6=h2dba641_0
13
+ - libgcc=14.2.0=h77fa898_1
14
+ - libgcc-ng=14.2.0=h69a702a_1
15
+ - libgomp=14.2.0=h77fa898_1
16
+ - liblzma=5.6.4=hb9d3cd8_0
17
+ - libnsl=2.0.1=hd590300_0
18
+ - libsqlite=3.49.1=hee588c1_1
19
+ - libuuid=2.38.1=h0b41bf4_0
20
+ - libxcrypt=4.4.36=hd590300_1
21
+ - libzlib=1.3.1=hb9d3cd8_2
22
+ - ncurses=6.5=h2d0b736_3
23
+ - openssl=3.4.1=h7b32b05_0
24
+ - pip=25.0.1=pyh8b19718_0
25
+ - python=3.11.11=h9e4cc4f_1_cpython
26
+ - readline=8.2=h8228510_1
27
+ - setuptools=75.8.0=pyhff2d567_0
28
+ - tk=8.6.13=noxft_h4845f30_101
29
+ - wheel=0.45.1=pyhd8ed1ab_1
30
+ - pip:
31
+ - accelerate==0.26.0
32
+ - aiofiles==23.2.1
33
+ - aiohappyeyeballs==2.4.6
34
+ - aiohttp==3.11.12
35
+ - aiosignal==1.3.2
36
+ - altair==5.5.0
37
+ - anyio==4.8.0
38
+ - attrs==25.1.0
39
+ - certifi==2025.1.31
40
+ - charset-normalizer==3.4.1
41
+ - click==8.1.8
42
+ - contourpy==1.3.1
43
+ - cycler==0.12.1
44
+ - datasets==2.15.0
45
+ - deepspeed==0.14.5
46
+ - dill==0.3.7
47
+ - einops==0.8.1
48
+ - fastapi==0.95.1
49
+ - ffmpy==0.5.0
50
+ - filelock==3.17.0
51
+ - fire==0.7.0
52
+ - flash-attn==2.7.1.post1
53
+ - fonttools==4.56.0
54
+ - frozenlist==1.5.0
55
+ - fsspec==2023.10.0
56
+ - gputil==1.4.0
57
+ - gradio==3.38.0
58
+ - gradio-client==1.7.1
59
+ - h11==0.14.0
60
+ - hjson==3.1.0
61
+ - httpcore==1.0.7
62
+ - httpx==0.28.1
63
+ - huggingface-hub==0.29.1
64
+ - idna==3.10
65
+ - jieba==0.42.1
66
+ - jinja2==3.1.5
67
+ - joblib==1.4.2
68
+ - jsonschema==4.23.0
69
+ - jsonschema-specifications==2024.10.1
70
+ - kiwisolver==1.4.8
71
+ - linkify-it-py==2.0.3
72
+ - markdown-it-py==2.2.0
73
+ - markupsafe==2.1.5
74
+ - matplotlib==3.10.0
75
+ - mdit-py-plugins==0.3.3
76
+ - mdurl==0.1.2
77
+ - mpmath==1.3.0
78
+ - multidict==6.1.0
79
+ - multiprocess==0.70.15
80
+ - narwhals==1.27.1
81
+ - networkx==3.4.2
82
+ - ninja==1.11.1.3
83
+ - nltk==3.9.1
84
+ - numpy==1.26.4
85
+ - nvidia-cublas-cu12==12.1.3.1
86
+ - nvidia-cuda-cupti-cu12==12.1.105
87
+ - nvidia-cuda-nvrtc-cu12==12.1.105
88
+ - nvidia-cuda-runtime-cu12==12.1.105
89
+ - nvidia-cudnn-cu12==8.9.2.26
90
+ - nvidia-cufft-cu12==11.0.2.54
91
+ - nvidia-curand-cu12==10.3.2.106
92
+ - nvidia-cusolver-cu12==11.4.5.107
93
+ - nvidia-cusparse-cu12==12.1.0.106
94
+ - nvidia-ml-py==12.570.86
95
+ - nvidia-nccl-cu12==2.18.1
96
+ - nvidia-nvjitlink-cu12==12.8.61
97
+ - nvidia-nvtx-cu12==12.1.105
98
+ - orjson==3.10.15
99
+ - packaging==24.2
100
+ - pandas==2.2.3
101
+ - peft==0.6.1
102
+ - pillow==10.4.0
103
+ - propcache==0.3.0
104
+ - protobuf==5.29.3
105
+ - psutil==7.0.0
106
+ - py-cpuinfo==9.0.0
107
+ - pyarrow==19.0.1
108
+ - pyarrow-hotfix==0.6
109
+ - pydantic==1.10.11
110
+ - pydub==0.25.1
111
+ - pyparsing==3.2.1
112
+ - python-dateutil==2.9.0.post0
113
+ - python-multipart==0.0.20
114
+ - pytz==2025.1
115
+ - pyyaml==6.0.2
116
+ - referencing==0.36.2
117
+ - regex==2024.11.6
118
+ - requests==2.32.3
119
+ - rouge-chinese==1.0.3
120
+ - rpds-py==0.23.1
121
+ - safetensors==0.5.2
122
+ - scipy==1.15.2
123
+ - semantic-version==2.10.0
124
+ - sentencepiece==0.2.0
125
+ - six==1.17.0
126
+ - sniffio==1.3.1
127
+ - sse-starlette==2.1.3
128
+ - starlette==0.26.1
129
+ - sympy==1.13.3
130
+ - termcolor==2.5.0
131
+ - tiktoken==0.9.0
132
+ - tokenizers==0.20.3
133
+ - torch==2.1.0
134
+ - tqdm==4.67.1
135
+ - transformers==4.46.0
136
+ - triton==2.1.0
137
+ - typing-extensions==4.12.2
138
+ - tzdata==2025.1
139
+ - uc-micro-py==1.0.3
140
+ - urllib3==2.3.0
141
+ - uvicorn==0.34.0
142
+ - websockets==11.0.3
143
+ - xxhash==3.5.0
144
+ - yarl==1.18.3
145
+ prefix: /opt/aps/workdir/miniforge3/envs/train
deep_search/sft/train_requirements.txt ADDED
@@ -0,0 +1,116 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ torch==2.1.0
2
+ transformers==4.46.0
3
+ accelerate==0.26.0
4
+ aiofiles==23.2.1
5
+ aiohappyeyeballs==2.4.6
6
+ aiohttp==3.11.12
7
+ aiosignal==1.3.2
8
+ altair==5.5.0
9
+ anyio==4.8.0
10
+ attrs==25.1.0
11
+ certifi==2025.1.31
12
+ charset-normalizer==3.4.1
13
+ click==8.1.8
14
+ contourpy==1.3.1
15
+ cycler==0.12.1
16
+ datasets==2.15.0
17
+ deepspeed==0.14.5
18
+ dill==0.3.7
19
+ einops==0.8.1
20
+ fastapi==0.95.1
21
+ ffmpy==0.5.0
22
+ filelock==3.17.0
23
+ fire==0.7.0
24
+ # flash-attn==2.7.1.post4
25
+ fonttools==4.56.0
26
+ frozenlist==1.5.0
27
+ fsspec==2023.10.0
28
+ gputil==1.4.0
29
+ gradio==3.38.0
30
+ gradio-client==1.7.1
31
+ h11==0.14.0
32
+ hjson==3.1.0
33
+ httpcore==1.0.7
34
+ httpx==0.28.1
35
+ huggingface-hub==0.29.1
36
+ idna==3.10
37
+ jieba==0.42.1
38
+ jinja2==3.1.5
39
+ joblib==1.4.2
40
+ jsonschema==4.23.0
41
+ jsonschema-specifications==2024.10.1
42
+ kiwisolver==1.4.8
43
+ linkify-it-py==2.0.3
44
+ markdown-it-py==2.2.0
45
+ markupsafe==2.1.5
46
+ matplotlib==3.10.0
47
+ mdit-py-plugins==0.3.3
48
+ mdurl==0.1.2
49
+ mpmath==1.3.0
50
+ multidict==6.1.0
51
+ multiprocess==0.70.15
52
+ narwhals==1.27.1
53
+ networkx==3.4.2
54
+ ninja==1.11.1.3
55
+ nltk==3.9.1
56
+ numpy==1.26.4
57
+ nvidia-cublas-cu12==12.1.3.1
58
+ nvidia-cuda-cupti-cu12==12.1.105
59
+ nvidia-cuda-nvrtc-cu12==12.1.105
60
+ nvidia-cuda-runtime-cu12==12.1.105
61
+ nvidia-cudnn-cu12==8.9.2.26
62
+ nvidia-cufft-cu12==11.0.2.54
63
+ nvidia-curand-cu12==10.3.2.106
64
+ nvidia-cusolver-cu12==11.4.5.107
65
+ nvidia-cusparse-cu12==12.1.0.106
66
+ nvidia-ml-py==12.570.86
67
+ nvidia-nccl-cu12==2.18.1
68
+ nvidia-nvjitlink-cu12==12.8.61
69
+ nvidia-nvtx-cu12==12.1.105
70
+ orjson==3.10.15
71
+ packaging==24.2
72
+ pandas==2.2.3
73
+ peft==0.6.1
74
+ pillow==10.4.0
75
+ propcache==0.3.0
76
+ protobuf==5.29.3
77
+ psutil==7.0.0
78
+ py-cpuinfo==9.0.0
79
+ pyarrow==19.0.1
80
+ pyarrow-hotfix==0.6
81
+ pydantic==1.10.11
82
+ pydub==0.25.1
83
+ pyparsing==3.2.1
84
+ python-dateutil==2.9.0.post0
85
+ python-multipart==0.0.20
86
+ pytz==2025.1
87
+ pyyaml==6.0.2
88
+ referencing==0.36.2
89
+ regex==2024.11.6
90
+ requests==2.32.3
91
+ rouge-chinese==1.0.3
92
+ rpds-py==0.23.1
93
+ safetensors==0.5.2
94
+ scipy==1.15.2
95
+ semantic-version==2.10.0
96
+ sentencepiece==0.2.0
97
+ six==1.17.0
98
+ sniffio==1.3.1
99
+ sse-starlette==2.1.3
100
+ starlette==0.26.1
101
+ sympy==1.13.3
102
+ termcolor==2.5.0
103
+ tiktoken==0.9.0
104
+ tokenizers==0.20.3
105
+
106
+ tqdm==4.67.1
107
+
108
+ triton==2.1.0
109
+ typing-extensions==4.12.2
110
+ tzdata==2025.1
111
+ uc-micro-py==1.0.3
112
+ urllib3==2.3.0
113
+ uvicorn==0.34.0
114
+ websockets==11.0.3
115
+ xxhash==3.5.0
116
+ yarl==1.18.3
deep_search/sft/wait_eval_1.py ADDED
@@ -0,0 +1,167 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import os
2
+ from time import sleep
3
+ import datetime
4
+ import subprocess
5
+ import json
6
+ from queue import Queue
7
+ import GPUtil
8
+ import sys
9
+
10
+ wait_cnt = 0
11
+ SLEEP_TIME = 120
12
+ NEED_REVERSE = False
13
+ eppt = None
14
+ NEED_ATLAS = False
15
+ only_rank = False
16
+ commands = Queue()
17
+
18
+ # print(f"wait 120s")
19
+ # sleep(120)
20
+
21
+ def get_free_gpu(threshold=0.95):
22
+ # 获取所有可用的GPU设备
23
+ gpus = GPUtil.getGPUs()
24
+ # 筛选出显存空闲率高于阈值的GPU
25
+ available_gpus = [gpu.id for gpu in gpus if gpu.memoryFree > 80000]
26
+ print(f"available_gpus: {available_gpus}")
27
+ return available_gpus
28
+
29
+ def run(commands):
30
+ gpus = get_free_gpu()
31
+ if NEED_REVERSE:
32
+ gpus.reverse()
33
+ while not commands.empty() and len(gpus) >= 2:
34
+ command = commands.get()
35
+ print(f"The following command is about to run:\n{command.format(device=f'{gpus[0]},{gpus[1]}')}")
36
+ subprocess.Popen(command.format(device=f'{gpus[0]},{gpus[1]}'), shell=True, start_new_session=True)
37
+ gpus = gpus[2:]
38
+ sleep(SLEEP_TIME)
39
+ return commands
40
+
41
+ def main():
42
+ global commands, wait_cnt, eppt, end_ckpt, only_rank
43
+ eppt = ckpt
44
+ math_datasets = [
45
+ 'AIME24',
46
+ ]
47
+ qa_datasets = [
48
+ 'aime', 'hotpotqa','2wiki', 'musique', "amc", "math500"
49
+ ]
50
+
51
+ ckpt_dir = f"{ckpt}"
52
+
53
+ all_items = os.listdir(ckpt_dir)
54
+ step_list = []
55
+ for item in all_items:
56
+ if item.startswith("checkpoint-"):
57
+ try:
58
+ step = int(item.split("-")[1])
59
+ step_list.append(step)
60
+ except ValueError:
61
+ continue
62
+
63
+ step_list.sort()
64
+
65
+ for step in step_list:
66
+ checkpoint_dir = f"{ckpt}/checkpoint-{step}"
67
+ export_dir = checkpoint_dir
68
+ wait_cnt = 0
69
+ print(f"Waiting for checkpoint ({checkpoint_dir}) to exist...")
70
+
71
+ while not os.path.exists(os.path.join(checkpoint_dir, "special_tokens_map.json")):
72
+ commands = run(commands)
73
+ wait_cnt += 1
74
+ print(f"Already waiting {datetime.timedelta(seconds=SLEEP_TIME*wait_cnt)}.")
75
+ if wait_cnt > (86400 // SLEEP_TIME):
76
+ clear()
77
+ raise Exception("There have been no new checkpoints for no less than a day, and the program is about to automatically exit.")
78
+ # python3 = "/opt/aps/workdir/miniforge3/envs/search_o1/bin/python3"
79
+ python3 = "/opt/aps/workdir/miniforge3/envs/search_o1/bin/python"
80
+ # /opt/aps/workdir/miniforge3/envs/search_o1/bin/python
81
+ print("The checkpoint exists. Waiting for running...")
82
+ for dataset in qa_datasets:
83
+ base_name = f"{os.path.basename(ckpt)}/{step}/{dataset}"
84
+ dir = "/opt/aps/workdir/sunshuang/search_o1"
85
+ # cache_path = f"{dir}/cache_reason_two_model/eval_reason_two_model/qwen-instruct-32B/{base_name}"
86
+ cache_path = f"{dir}/cache_reason_two_model/eval_reason_two_model"
87
+ output_path = f"{dir}/outputs_reason_two_model/eval_reason_two_model/qwen-instruct-32B/{base_name}"
88
+ log_path = f"{dir}/logs/eval_reason_two_model/qwen-instruct-32B/{base_name}"
89
+ os.makedirs(cache_path, exist_ok=True)
90
+ os.makedirs(output_path, exist_ok=True)
91
+ os.makedirs(log_path, exist_ok=True)
92
+ contents = os.listdir(output_path)
93
+
94
+ subset_num = 100
95
+ if dataset == "math500":
96
+ subset_num = 500
97
+
98
+
99
+ if not contents:
100
+ command = (
101
+ "export http_proxy=http://127.0.0.1:7880 && "
102
+ "export https_proxy=http://127.0.0.1:7880 && "
103
+ "cd /opt/aps/workdir/sunshuang/search_o1/ && "
104
+ "CUDA_VISIBLE_DEVICES={device} "
105
+ f"nohup {python3} -u scripts/reason_two_model_5.py "
106
+ f"--dataset_name {dataset} "
107
+ f"--cache_dir_base {cache_path} "
108
+ f"--output_dir_base {output_path} "
109
+ "--split test "
110
+ "--max_search_limit 5 "
111
+ "--max_turn 10 "
112
+ "--top_k 5 "
113
+ "--max_doc_len 3000 "
114
+ f"--subset_num {subset_num} "
115
+ f"--model_path {checkpoint_dir} "
116
+ "--model_doc_reason_path \"/capacity/userdata/models/Qwen2.5-32B-Instruct\" "
117
+ "--jina_api_key \"jina_04d684ee4cc54d2ebe7c43bb7ad4aff0qlkdZGwm14NFBtm5BDkgK9KNf6vQ\" "
118
+ "--bing_subscription_key \"cb0d28279a826d7e5cf22d71f683c77ffd4ba27d\" "
119
+ "--bing_endpoint \"https://google.serper.dev/search\" "
120
+ "--openai_api_base \"http://localhost:8001/v1\" "
121
+ f"> {log_path}/inf.log 2>&1 &"
122
+ )
123
+ commands.put(command)
124
+ else:
125
+ print(f"skip evaluated model: {base_name}")
126
+
127
+ continue
128
+
129
+ for dataset in math_datasets:
130
+ dir = "/opt/aps/workdir/math/evaluation"
131
+ base_name = f"{os.path.basename(ckpt)}/{step}/{dataset}"
132
+ output_path = f"{dir}/outputs/{base_name}"
133
+ log_path = f"{dir}/logs/{base_name}"
134
+ os.makedirs(output_path, exist_ok=True)
135
+ os.makedirs(log_path, exist_ok=True)
136
+ contents = os.listdir(output_path)
137
+ if not contents:
138
+ command = (
139
+ "cd /opt/aps/workdir/math/evaluation && "
140
+ "CUDA_VISIBLE_DEVICES={device} "
141
+ f"nohup {python3} -u run_eval_32b.py "
142
+ f"--data_name {dataset} "
143
+ f"--target_path {output_path} "
144
+ f"--model_name_or_path {checkpoint_dir} "
145
+ "--prompt v4 "
146
+ "--max_tokens 8192 "
147
+ "--paralle_size 1 "
148
+ f"> {log_path}/inf.log 2>&1 &"
149
+ )
150
+ commands.put(command)
151
+ else:
152
+ print(f"skip evaluated model: {base_name}")
153
+
154
+ def clear():
155
+ print("All checkpoints exist. Wait for runing...")
156
+ global commands
157
+ while not commands.empty():
158
+ # sleep(SLEEP_TIME)
159
+ commands = run(commands)
160
+ print("Wish me good luck!")
161
+
162
+
163
+ if __name__ == '__main__':
164
+ # print(sys.argv[1])
165
+ ckpt = sys.argv[1]
166
+ main()
167
+ clear()
deep_search/sft/wait_eval_use_one_model_for_ckpt.py ADDED
@@ -0,0 +1,178 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import os
2
+ from time import sleep
3
+ import datetime
4
+ import subprocess
5
+ import json
6
+ from queue import Queue
7
+ import GPUtil
8
+ import sys
9
+ import random
10
+
11
+ wait_cnt = 0
12
+ SLEEP_TIME = 120
13
+ NEED_REVERSE = False
14
+ eppt = None
15
+ NEED_ATLAS = False
16
+ only_rank = False
17
+ commands = Queue()
18
+
19
+ # print(f"wait 120s")
20
+ # sleep(120)
21
+
22
+ def get_free_gpu(threshold=0.95):
23
+ # 获取所有可用的GPU设备
24
+ gpus = GPUtil.getGPUs()
25
+ # 筛选出显存空闲率高于阈值的GPU
26
+ available_gpus = [gpu.id for gpu in gpus if gpu.memoryFree > 80000 ]
27
+ print(f"available_gpus: {available_gpus}")
28
+ return available_gpus
29
+
30
+ def run(commands):
31
+ gpus = get_free_gpu()
32
+ if NEED_REVERSE:
33
+ gpus.reverse()
34
+ # random.shuffle(gpus)
35
+ # print(f"available_gpus: {gpus}")
36
+ while not commands.empty() and len(gpus) >= 2:
37
+ command = commands.get()
38
+ print(f"The following command is about to run:\n{command.format(device=f'{gpus[0]},{gpus[1]}')}")
39
+ subprocess.Popen(command.format(device=f'{gpus[0]},{gpus[1]}'), shell=True, start_new_session=True)
40
+ gpus = gpus[2:]
41
+ sleep(SLEEP_TIME)
42
+ return commands
43
+
44
+ def main():
45
+ global commands, wait_cnt, eppt, end_ckpt, only_rank
46
+ eppt = ckpt
47
+ math_datasets = [
48
+ 'AIME24',
49
+ ]
50
+ # qa_datasets = [
51
+ # 'aime', 'hotpotqa','2wiki', 'musique', "amc", "math500"
52
+ # ]
53
+ # qa_datasets = [
54
+ # 'simpleqa', 'hotpotqa'
55
+ # ]
56
+ qa_datasets = [
57
+ 'eval_old_500'
58
+ ]
59
+ ckpt_dir = f"{ckpt}"
60
+
61
+ all_items = os.listdir(ckpt_dir)
62
+ step_list = []
63
+ for item in all_items:
64
+ if item.startswith("checkpoint-"):
65
+ try:
66
+ step = int(item.split("-")[1])
67
+ step_list.append(step)
68
+ except ValueError:
69
+ continue
70
+
71
+ step_list.sort(reverse=True) # 从大到小排序
72
+ # print(f"step_list: {step_list}")
73
+ # # step_list = step_list[:2] # 只取最新的checkpoint
74
+ print(f"step_list: {step_list}")
75
+ step_list = step_list[:4]
76
+ # step_list = ['55']
77
+ print(f"step_list: {step_list}")
78
+ for step in step_list:
79
+ checkpoint_dir = f"{ckpt}/checkpoint-{step}"
80
+ export_dir = checkpoint_dir
81
+ wait_cnt = 0
82
+ print(f"Waiting for checkpoint ({checkpoint_dir}) to exist...")
83
+
84
+ while not os.path.exists(os.path.join(checkpoint_dir, "special_tokens_map.json")):
85
+ commands = run(commands)
86
+ wait_cnt += 1
87
+ print(f"Already waiting {datetime.timedelta(seconds=SLEEP_TIME*wait_cnt)}.")
88
+ if wait_cnt > (86400 // SLEEP_TIME):
89
+ clear()
90
+ raise Exception("There have been no new checkpoints for no less than a day, and the program is about to automatically exit.")
91
+ # python3 = "/opt/aps/workdir/miniforge3/envs/search_o1/bin/python3"
92
+ python3 = "/opt/aps/workdir/search_o1/bin/python3"
93
+ # /opt/aps/workdir/miniforge3/envs/search_o1/bin/python
94
+ print("The checkpoint exists. Waiting for running...")
95
+ for dataset in qa_datasets:
96
+ base_name = f"{os.path.basename(ckpt)}/{step}/{dataset}"
97
+ dir = "/opt/aps/workdir/sunshuang/deep_search/search_o1"
98
+ # cache_path = f"{dir}/cache_reason_two_model/eval_reason_two_model/qwen-instruct-32B/{base_name}"
99
+ cache_path = f"{dir}/cache_eval_sum_all_webpage"
100
+ output_path = f"{dir}/output/output_eval/{base_name}"
101
+ log_path = f"{dir}/logs/log_eval/{base_name}"
102
+ os.makedirs(cache_path, exist_ok=True)
103
+ os.makedirs(output_path, exist_ok=True)
104
+ os.makedirs(log_path, exist_ok=True)
105
+ contents = os.listdir(output_path)
106
+
107
+ # subset_num = 500
108
+ # if dataset == "math500":
109
+ # subset_num = 500
110
+
111
+
112
+ if not contents:
113
+ command = (
114
+ "export https_proxy=http://127.0.0.1:7890 && "
115
+ "export http_proxy=http://127.0.0.1:7890 && "
116
+ "export all_proxy=socks5://127.0.0.1:7891 && "
117
+ "cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && "
118
+ "CUDA_VISIBLE_DEVICES={device} "
119
+ f"nohup {python3} -u scripts/search_o1_sum_all_webpage.py "
120
+ f"--dataset_name {dataset} "
121
+ f"--cache_dir_base {cache_path} "
122
+ f"--output_dir_base {output_path} "
123
+ "--split test "
124
+ "--max_search_limit 10 "
125
+ "--max_turn 10 "
126
+ "--top_k 10 "
127
+ "--max_doc_len 5000 "
128
+ f"--model_path {checkpoint_dir} "
129
+ "--bing_subscription_key \"cb0d28279a826d7e5cf22d71f683c77ffd4ba27d\" "
130
+ "--bing_endpoint \"https://google.serper.dev/search\" "
131
+ f"> {log_path}/inf.log 2>&1 &"
132
+ )
133
+ print(f"command: {command}")
134
+ commands.put(command)
135
+ else:
136
+ print(f"skip evaluated model: {base_name}")
137
+
138
+ # continue
139
+
140
+ # for dataset in math_datasets:
141
+ # dir = "/opt/aps/workdir/math/evaluation"
142
+ # base_name = f"{os.path.basename(ckpt)}/{step}/{dataset}"
143
+ # output_path = f"{dir}/outputs_maxtoken_20000/{base_name}"
144
+ # log_path = f"{dir}/logs_maxtoken_20000/{base_name}"
145
+ # os.makedirs(output_path, exist_ok=True)
146
+ # os.makedirs(log_path, exist_ok=True)
147
+ # contents = os.listdir(output_path)
148
+ # if not contents:
149
+ # command = (
150
+ # "cd /opt/aps/workdir/math/evaluation && "
151
+ # "CUDA_VISIBLE_DEVICES={device} "
152
+ # f"nohup {python3} -u run_eval_32b.py "
153
+ # f"--data_name {dataset} "
154
+ # f"--target_path {output_path} "
155
+ # f"--model_name_or_path {checkpoint_dir} "
156
+ # "--prompt v4 "
157
+ # "--max_tokens 20000 "
158
+ # "--paralle_size 2 "
159
+ # f"> {log_path}/inf.log 2>&1 &"
160
+ # )
161
+ # commands.put(command)
162
+ # else:
163
+ # print(f"skip evaluated model: {base_name}")
164
+
165
+ def clear():
166
+ print("All checkpoints exist. Wait for runing...")
167
+ global commands
168
+ while not commands.empty():
169
+ # sleep(SLEEP_TIME)
170
+ commands = run(commands)
171
+ print("Wish me good luck!")
172
+
173
+
174
+ if __name__ == '__main__':
175
+ # print(sys.argv[1])
176
+ ckpt = sys.argv[1]
177
+ main()
178
+ clear()
deep_search/sft/wait_eval_use_one_model_for_ckpt_two_model.py ADDED
@@ -0,0 +1,178 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import os
2
+ from time import sleep
3
+ import datetime
4
+ import subprocess
5
+ import json
6
+ from queue import Queue
7
+ import GPUtil
8
+ import sys
9
+ import random
10
+
11
+ wait_cnt = 0
12
+ SLEEP_TIME = 120
13
+ NEED_REVERSE = False
14
+ eppt = None
15
+ NEED_ATLAS = False
16
+ only_rank = False
17
+ commands = Queue()
18
+
19
+ # print(f"wait 120s")
20
+ # sleep(120)
21
+
22
+ def get_free_gpu(threshold=0.95):
23
+ # 获取所有可用的GPU设备
24
+ gpus = GPUtil.getGPUs()
25
+ # 筛选出显存空闲率高于阈值的GPU
26
+ available_gpus = [gpu.id for gpu in gpus if gpu.memoryFree > 80000 ]
27
+ print(f"available_gpus: {available_gpus}")
28
+ return available_gpus
29
+
30
+ def run(commands):
31
+ gpus = get_free_gpu()
32
+ if NEED_REVERSE:
33
+ gpus.reverse()
34
+ # random.shuffle(gpus)
35
+ # print(f"available_gpus: {gpus}")
36
+ while not commands.empty() and len(gpus) >= 2:
37
+ command = commands.get()
38
+ print(f"The following command is about to run:\n{command.format(device=f'{gpus[0]},{gpus[1]}')}")
39
+ subprocess.Popen(command.format(device=f'{gpus[0]},{gpus[1]}'), shell=True, start_new_session=True)
40
+ gpus = gpus[2:]
41
+ sleep(SLEEP_TIME)
42
+ return commands
43
+
44
+ def main():
45
+ global commands, wait_cnt, eppt, end_ckpt, only_rank
46
+ eppt = ckpt
47
+ math_datasets = [
48
+ 'AIME24',
49
+ ]
50
+ # qa_datasets = [
51
+ # 'aime', 'hotpotqa','2wiki', 'musique', "amc", "math500"
52
+ # ]
53
+ # qa_datasets = [
54
+ # 'simpleqa', 'hotpotqa'
55
+ # ]
56
+ qa_datasets = [
57
+ 'eval', 'gaia'
58
+ ]
59
+ ckpt_dir = f"{ckpt}"
60
+
61
+ all_items = os.listdir(ckpt_dir)
62
+ step_list = []
63
+ for item in all_items:
64
+ if item.startswith("checkpoint-"):
65
+ try:
66
+ step = int(item.split("-")[1])
67
+ step_list.append(step)
68
+ except ValueError:
69
+ continue
70
+
71
+ step_list.sort(reverse=True) # 从大到小排序
72
+ # print(f"step_list: {step_list}")
73
+ # # step_list = step_list[:2] # 只取最新的checkpoint
74
+ print(f"step_list: {step_list}")
75
+ # step_list = step_list[:4]
76
+ # step_list = ['55']
77
+ print(f"step_list: {step_list}")
78
+ for step in step_list:
79
+ checkpoint_dir = f"{ckpt}/checkpoint-{step}"
80
+ export_dir = checkpoint_dir
81
+ wait_cnt = 0
82
+ print(f"Waiting for checkpoint ({checkpoint_dir}) to exist...")
83
+
84
+ while not os.path.exists(os.path.join(checkpoint_dir, "special_tokens_map.json")):
85
+ commands = run(commands)
86
+ wait_cnt += 1
87
+ print(f"Already waiting {datetime.timedelta(seconds=SLEEP_TIME*wait_cnt)}.")
88
+ if wait_cnt > (86400 // SLEEP_TIME):
89
+ clear()
90
+ raise Exception("There have been no new checkpoints for no less than a day, and the program is about to automatically exit.")
91
+ # python3 = "/opt/aps/workdir/miniforge3/envs/search_o1/bin/python3"
92
+ python3 = "/opt/aps/workdir/search_o1/bin/python3"
93
+ # /opt/aps/workdir/miniforge3/envs/search_o1/bin/python
94
+ print("The checkpoint exists. Waiting for running...")
95
+ for dataset in qa_datasets:
96
+ base_name = f"{os.path.basename(ckpt)}/{step}/{dataset}"
97
+ dir = "/opt/aps/workdir/sunshuang/deep_search/search_o1"
98
+ # cache_path = f"{dir}/cache_reason_two_model/eval_reason_two_model/qwen-instruct-32B/{base_name}"
99
+ cache_path = f"{dir}/cache_eval_sum_all_webpage_1w"
100
+ output_path = f"{dir}/output/output_eval/{base_name}"
101
+ log_path = f"{dir}/logs/log_eval/{base_name}"
102
+ os.makedirs(cache_path, exist_ok=True)
103
+ os.makedirs(output_path, exist_ok=True)
104
+ os.makedirs(log_path, exist_ok=True)
105
+ contents = os.listdir(output_path)
106
+
107
+ # subset_num = 500
108
+ # if dataset == "math500":
109
+ # subset_num = 500
110
+
111
+
112
+ if not contents:
113
+ command = (
114
+ "export https_proxy=http://127.0.0.1:7890 && "
115
+ "export http_proxy=http://127.0.0.1:7890 && "
116
+ "export all_proxy=socks5://127.0.0.1:7891 && "
117
+ "cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && "
118
+ "CUDA_VISIBLE_DEVICES={device} "
119
+ f"nohup {python3} -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm.py "
120
+ f"--dataset_name {dataset} "
121
+ f"--cache_dir_base {cache_path} "
122
+ f"--output_dir_base {output_path} "
123
+ "--split test "
124
+ "--max_search_limit 10 "
125
+ "--max_turn 10 "
126
+ "--top_k 10 "
127
+ "--max_doc_len 5000 "
128
+ f"--model_path {checkpoint_dir} "
129
+ "--bing_subscription_key \"cb0d28279a826d7e5cf22d71f683c77ffd4ba27d\" "
130
+ "--bing_endpoint \"https://google.serper.dev/search\" "
131
+ f"> {log_path}/inf.log 2>&1 &"
132
+ )
133
+ print(f"command: {command}")
134
+ commands.put(command)
135
+ else:
136
+ print(f"skip evaluated model: {base_name}")
137
+
138
+ # continue
139
+
140
+ # for dataset in math_datasets:
141
+ # dir = "/opt/aps/workdir/math/evaluation"
142
+ # base_name = f"{os.path.basename(ckpt)}/{step}/{dataset}"
143
+ # output_path = f"{dir}/outputs_maxtoken_20000/{base_name}"
144
+ # log_path = f"{dir}/logs_maxtoken_20000/{base_name}"
145
+ # os.makedirs(output_path, exist_ok=True)
146
+ # os.makedirs(log_path, exist_ok=True)
147
+ # contents = os.listdir(output_path)
148
+ # if not contents:
149
+ # command = (
150
+ # "cd /opt/aps/workdir/math/evaluation && "
151
+ # "CUDA_VISIBLE_DEVICES={device} "
152
+ # f"nohup {python3} -u run_eval_32b.py "
153
+ # f"--data_name {dataset} "
154
+ # f"--target_path {output_path} "
155
+ # f"--model_name_or_path {checkpoint_dir} "
156
+ # "--prompt v4 "
157
+ # "--max_tokens 20000 "
158
+ # "--paralle_size 2 "
159
+ # f"> {log_path}/inf.log 2>&1 &"
160
+ # )
161
+ # commands.put(command)
162
+ # else:
163
+ # print(f"skip evaluated model: {base_name}")
164
+
165
+ def clear():
166
+ print("All checkpoints exist. Wait for runing...")
167
+ global commands
168
+ while not commands.empty():
169
+ # sleep(SLEEP_TIME)
170
+ commands = run(commands)
171
+ print("Wish me good luck!")
172
+
173
+
174
+ if __name__ == '__main__':
175
+ # print(sys.argv[1])
176
+ ckpt = sys.argv[1]
177
+ main()
178
+ clear()
deep_search/sft/wait_eval_use_one_model_for_ckpt_two_model_dpsk.py ADDED
@@ -0,0 +1,178 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import os
2
+ from time import sleep
3
+ import datetime
4
+ import subprocess
5
+ import json
6
+ from queue import Queue
7
+ import GPUtil
8
+ import sys
9
+ import random
10
+
11
+ wait_cnt = 0
12
+ SLEEP_TIME = 120
13
+ NEED_REVERSE = False
14
+ eppt = None
15
+ NEED_ATLAS = False
16
+ only_rank = False
17
+ commands = Queue()
18
+
19
+ # print(f"wait 120s")
20
+ # sleep(120)
21
+
22
+ def get_free_gpu(threshold=0.95):
23
+ # 获取所有可用的GPU设备
24
+ gpus = GPUtil.getGPUs()
25
+ # 筛选出显存空闲率高于阈值的GPU
26
+ available_gpus = [gpu.id for gpu in gpus if gpu.memoryFree > 80000 ]
27
+ print(f"available_gpus: {available_gpus}")
28
+ return available_gpus
29
+
30
+ def run(commands):
31
+ gpus = get_free_gpu()
32
+ if NEED_REVERSE:
33
+ gpus.reverse()
34
+ # random.shuffle(gpus)
35
+ # print(f"available_gpus: {gpus}")
36
+ while not commands.empty() and len(gpus) >= 2:
37
+ command = commands.get()
38
+ print(f"The following command is about to run:\n{command.format(device=f'{gpus[0]},{gpus[1]}')}")
39
+ subprocess.Popen(command.format(device=f'{gpus[0]},{gpus[1]}'), shell=True, start_new_session=True)
40
+ gpus = gpus[2:]
41
+ sleep(SLEEP_TIME)
42
+ return commands
43
+
44
+ def main():
45
+ global commands, wait_cnt, eppt, end_ckpt, only_rank
46
+ eppt = ckpt
47
+ math_datasets = [
48
+ 'AIME24',
49
+ ]
50
+ # qa_datasets = [
51
+ # 'aime', 'hotpotqa','2wiki', 'musique', "amc", "math500"
52
+ # ]
53
+ # qa_datasets = [
54
+ # 'simpleqa', 'hotpotqa'
55
+ # ]
56
+ qa_datasets = [
57
+ 'eval', 'gaia', 'aime'
58
+ ]
59
+ ckpt_dir = f"{ckpt}"
60
+
61
+ all_items = os.listdir(ckpt_dir)
62
+ step_list = []
63
+ for item in all_items:
64
+ if item.startswith("checkpoint-"):
65
+ try:
66
+ step = int(item.split("-")[1])
67
+ step_list.append(step)
68
+ except ValueError:
69
+ continue
70
+
71
+ step_list.sort(reverse=True) # 从大到小排序
72
+ # print(f"step_list: {step_list}")
73
+ # # step_list = step_list[:2] # 只取最新的checkpoint
74
+ print(f"step_list: {step_list}")
75
+ # step_list = step_list[:4]
76
+ # step_list = ['55']
77
+ print(f"step_list: {step_list}")
78
+ for step in step_list:
79
+ checkpoint_dir = f"{ckpt}/checkpoint-{step}"
80
+ export_dir = checkpoint_dir
81
+ wait_cnt = 0
82
+ print(f"Waiting for checkpoint ({checkpoint_dir}) to exist...")
83
+
84
+ while not os.path.exists(os.path.join(checkpoint_dir, "special_tokens_map.json")):
85
+ commands = run(commands)
86
+ wait_cnt += 1
87
+ print(f"Already waiting {datetime.timedelta(seconds=SLEEP_TIME*wait_cnt)}.")
88
+ if wait_cnt > (86400 // SLEEP_TIME):
89
+ clear()
90
+ raise Exception("There have been no new checkpoints for no less than a day, and the program is about to automatically exit.")
91
+ # python3 = "/opt/aps/workdir/miniforge3/envs/search_o1/bin/python3"
92
+ python3 = "/opt/aps/workdir/search_o1/bin/python3"
93
+ # /opt/aps/workdir/miniforge3/envs/search_o1/bin/python
94
+ print("The checkpoint exists. Waiting for running...")
95
+ for dataset in qa_datasets:
96
+ base_name = f"{os.path.basename(ckpt)}/{step}/{dataset}"
97
+ dir = "/opt/aps/workdir/sunshuang/deep_search/search_o1"
98
+ # cache_path = f"{dir}/cache_reason_two_model/eval_reason_two_model/qwen-instruct-32B/{base_name}"
99
+ cache_path = f"{dir}/cache_eval_sum_all_webpage_1w"
100
+ output_path = f"{dir}/output/output_eval/{base_name}"
101
+ log_path = f"{dir}/logs/log_eval/{base_name}"
102
+ os.makedirs(cache_path, exist_ok=True)
103
+ os.makedirs(output_path, exist_ok=True)
104
+ os.makedirs(log_path, exist_ok=True)
105
+ contents = os.listdir(output_path)
106
+
107
+ # subset_num = 500
108
+ # if dataset == "math500":
109
+ # subset_num = 500
110
+
111
+
112
+ if not contents:
113
+ command = (
114
+ "export https_proxy=http://127.0.0.1:7890 && "
115
+ "export http_proxy=http://127.0.0.1:7890 && "
116
+ "export all_proxy=socks5://127.0.0.1:7891 && "
117
+ "cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && "
118
+ "CUDA_VISIBLE_DEVICES={device} "
119
+ f"nohup {python3} -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_dpsk.py "
120
+ f"--dataset_name {dataset} "
121
+ f"--cache_dir_base {cache_path} "
122
+ f"--output_dir_base {output_path} "
123
+ "--split test "
124
+ "--max_search_limit 10 "
125
+ "--max_turn 10 "
126
+ "--top_k 10 "
127
+ "--max_doc_len 5000 "
128
+ f"--model_path {checkpoint_dir} "
129
+ "--bing_subscription_key \"cb0d28279a826d7e5cf22d71f683c77ffd4ba27d\" "
130
+ "--bing_endpoint \"https://google.serper.dev/search\" "
131
+ f"> {log_path}/inf.log 2>&1 &"
132
+ )
133
+ print(f"command: {command}")
134
+ commands.put(command)
135
+ else:
136
+ print(f"skip evaluated model: {base_name}")
137
+
138
+ # continue
139
+
140
+ # for dataset in math_datasets:
141
+ # dir = "/opt/aps/workdir/math/evaluation"
142
+ # base_name = f"{os.path.basename(ckpt)}/{step}/{dataset}"
143
+ # output_path = f"{dir}/outputs_maxtoken_20000/{base_name}"
144
+ # log_path = f"{dir}/logs_maxtoken_20000/{base_name}"
145
+ # os.makedirs(output_path, exist_ok=True)
146
+ # os.makedirs(log_path, exist_ok=True)
147
+ # contents = os.listdir(output_path)
148
+ # if not contents:
149
+ # command = (
150
+ # "cd /opt/aps/workdir/math/evaluation && "
151
+ # "CUDA_VISIBLE_DEVICES={device} "
152
+ # f"nohup {python3} -u run_eval_32b.py "
153
+ # f"--data_name {dataset} "
154
+ # f"--target_path {output_path} "
155
+ # f"--model_name_or_path {checkpoint_dir} "
156
+ # "--prompt v4 "
157
+ # "--max_tokens 20000 "
158
+ # "--paralle_size 2 "
159
+ # f"> {log_path}/inf.log 2>&1 &"
160
+ # )
161
+ # commands.put(command)
162
+ # else:
163
+ # print(f"skip evaluated model: {base_name}")
164
+
165
+ def clear():
166
+ print("All checkpoints exist. Wait for runing...")
167
+ global commands
168
+ while not commands.empty():
169
+ # sleep(SLEEP_TIME)
170
+ commands = run(commands)
171
+ print("Wish me good luck!")
172
+
173
+
174
+ if __name__ == '__main__':
175
+ # print(sys.argv[1])
176
+ ckpt = sys.argv[1]
177
+ main()
178
+ clear()
deep_search/sft/wait_eval_use_one_model_for_ckpt_two_model_qwq.py ADDED
@@ -0,0 +1,178 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import os
2
+ from time import sleep
3
+ import datetime
4
+ import subprocess
5
+ import json
6
+ from queue import Queue
7
+ import GPUtil
8
+ import sys
9
+ import random
10
+
11
+ wait_cnt = 0
12
+ SLEEP_TIME = 120
13
+ NEED_REVERSE = False
14
+ eppt = None
15
+ NEED_ATLAS = False
16
+ only_rank = False
17
+ commands = Queue()
18
+
19
+ # print(f"wait 120s")
20
+ # sleep(120)
21
+
22
+ def get_free_gpu(threshold=0.95):
23
+ # 获取所有可用的GPU设备
24
+ gpus = GPUtil.getGPUs()
25
+ # 筛选出显存空闲率高于阈值的GPU
26
+ available_gpus = [gpu.id for gpu in gpus if gpu.memoryFree > 80000 ]
27
+ print(f"available_gpus: {available_gpus}")
28
+ return available_gpus
29
+
30
+ def run(commands):
31
+ gpus = get_free_gpu()
32
+ if NEED_REVERSE:
33
+ gpus.reverse()
34
+ # random.shuffle(gpus)
35
+ # print(f"available_gpus: {gpus}")
36
+ while not commands.empty() and len(gpus) >= 2:
37
+ command = commands.get()
38
+ print(f"The following command is about to run:\n{command.format(device=f'{gpus[0]},{gpus[1]}')}")
39
+ subprocess.Popen(command.format(device=f'{gpus[0]},{gpus[1]}'), shell=True, start_new_session=True)
40
+ gpus = gpus[2:]
41
+ sleep(SLEEP_TIME)
42
+ return commands
43
+
44
+ def main():
45
+ global commands, wait_cnt, eppt, end_ckpt, only_rank
46
+ eppt = ckpt
47
+ math_datasets = [
48
+ 'AIME24',
49
+ ]
50
+ # qa_datasets = [
51
+ # 'aime', 'hotpotqa','2wiki', 'musique', "amc", "math500"
52
+ # ]
53
+ # qa_datasets = [
54
+ # 'simpleqa', 'hotpotqa'
55
+ # ]
56
+ qa_datasets = [
57
+ 'gaia', 'bamboogle'
58
+ ]
59
+ ckpt_dir = f"{ckpt}"
60
+
61
+ all_items = os.listdir(ckpt_dir)
62
+ step_list = []
63
+ for item in all_items:
64
+ if item.startswith("checkpoint-"):
65
+ try:
66
+ step = int(item.split("-")[1])
67
+ step_list.append(step)
68
+ except ValueError:
69
+ continue
70
+
71
+ step_list.sort(reverse=True) # 从大到小排序
72
+ # print(f"step_list: {step_list}")
73
+ # # step_list = step_list[:2] # 只取最新的checkpoint
74
+ print(f"step_list: {step_list}")
75
+ # step_list = step_list[:3]
76
+ # step_list = ['55']
77
+ # print(f"step_list: {step_list}")
78
+ for step in step_list:
79
+ checkpoint_dir = f"{ckpt}/checkpoint-{step}"
80
+ export_dir = checkpoint_dir
81
+ wait_cnt = 0
82
+ print(f"Waiting for checkpoint ({checkpoint_dir}) to exist...")
83
+
84
+ while not os.path.exists(os.path.join(checkpoint_dir, "special_tokens_map.json")):
85
+ commands = run(commands)
86
+ wait_cnt += 1
87
+ print(f"Already waiting {datetime.timedelta(seconds=SLEEP_TIME*wait_cnt)}.")
88
+ if wait_cnt > (86400 // SLEEP_TIME):
89
+ clear()
90
+ raise Exception("There have been no new checkpoints for no less than a day, and the program is about to automatically exit.")
91
+ # python3 = "/opt/aps/workdir/miniforge3/envs/search_o1/bin/python3"
92
+ python3 = "/opt/aps/workdir/search_o1/bin/python3"
93
+ # /opt/aps/workdir/miniforge3/envs/search_o1/bin/python
94
+ print("The checkpoint exists. Waiting for running...")
95
+ for dataset in qa_datasets:
96
+ base_name = f"{os.path.basename(ckpt)}/{step}/{dataset}"
97
+ dir = "/opt/aps/workdir/sunshuang/deep_search/search_o1"
98
+ # cache_path = f"{dir}/cache_reason_two_model/eval_reason_two_model/qwen-instruct-32B/{base_name}"
99
+ cache_path = f"{dir}/cache_eval_sum_all_webpage_1w"
100
+ output_path = f"{dir}/output/output_eval/{base_name}"
101
+ log_path = f"{dir}/logs/log_eval/{base_name}"
102
+ os.makedirs(cache_path, exist_ok=True)
103
+ os.makedirs(output_path, exist_ok=True)
104
+ os.makedirs(log_path, exist_ok=True)
105
+ contents = os.listdir(output_path)
106
+
107
+ # subset_num = 500
108
+ # if dataset == "math500":
109
+ # subset_num = 500
110
+
111
+
112
+ if not contents:
113
+ command = (
114
+ "export https_proxy=http://127.0.0.1:7890 && "
115
+ "export http_proxy=http://127.0.0.1:7890 && "
116
+ "export all_proxy=socks5://127.0.0.1:7891 && "
117
+ "cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && "
118
+ "CUDA_VISIBLE_DEVICES={device} "
119
+ f"nohup {python3} -u scripts/search_o1_sum_all_webpage_two_model_pool_vllm_qwq.py "
120
+ f"--dataset_name {dataset} "
121
+ f"--cache_dir_base {cache_path} "
122
+ f"--output_dir_base {output_path} "
123
+ "--split test "
124
+ "--max_search_limit 10 "
125
+ "--max_turn 10 "
126
+ "--top_k 10 "
127
+ "--max_doc_len 5000 "
128
+ f"--model_path {checkpoint_dir} "
129
+ "--bing_subscription_key \"cb0d28279a826d7e5cf22d71f683c77ffd4ba27d\" "
130
+ "--bing_endpoint \"https://google.serper.dev/search\" "
131
+ f"> {log_path}/inf.log 2>&1 &"
132
+ )
133
+ print(f"command: {command}")
134
+ commands.put(command)
135
+ else:
136
+ print(f"skip evaluated model: {base_name}")
137
+
138
+ # continue
139
+
140
+ # for dataset in math_datasets:
141
+ # dir = "/opt/aps/workdir/math/evaluation"
142
+ # base_name = f"{os.path.basename(ckpt)}/{step}/{dataset}"
143
+ # output_path = f"{dir}/outputs_maxtoken_20000/{base_name}"
144
+ # log_path = f"{dir}/logs_maxtoken_20000/{base_name}"
145
+ # os.makedirs(output_path, exist_ok=True)
146
+ # os.makedirs(log_path, exist_ok=True)
147
+ # contents = os.listdir(output_path)
148
+ # if not contents:
149
+ # command = (
150
+ # "cd /opt/aps/workdir/math/evaluation && "
151
+ # "CUDA_VISIBLE_DEVICES={device} "
152
+ # f"nohup {python3} -u run_eval_32b.py "
153
+ # f"--data_name {dataset} "
154
+ # f"--target_path {output_path} "
155
+ # f"--model_name_or_path {checkpoint_dir} "
156
+ # "--prompt v4 "
157
+ # "--max_tokens 20000 "
158
+ # "--paralle_size 2 "
159
+ # f"> {log_path}/inf.log 2>&1 &"
160
+ # )
161
+ # commands.put(command)
162
+ # else:
163
+ # print(f"skip evaluated model: {base_name}")
164
+
165
+ def clear():
166
+ print("All checkpoints exist. Wait for runing...")
167
+ global commands
168
+ while not commands.empty():
169
+ # sleep(SLEEP_TIME)
170
+ commands = run(commands)
171
+ print("Wish me good luck!")
172
+
173
+
174
+ if __name__ == '__main__':
175
+ # print(sys.argv[1])
176
+ ckpt = sys.argv[1]
177
+ main()
178
+ clear()
deep_search/sft/wait_eval_use_one_model_for_ckpt_worker0.py ADDED
@@ -0,0 +1,180 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import os
2
+ from time import sleep
3
+ import datetime
4
+ import subprocess
5
+ import json
6
+ from queue import Queue
7
+ import GPUtil
8
+ import sys
9
+ import random
10
+
11
+ wait_cnt = 0
12
+ SLEEP_TIME = 120
13
+ NEED_REVERSE = False
14
+ eppt = None
15
+ NEED_ATLAS = False
16
+ only_rank = False
17
+ commands = Queue()
18
+
19
+ # print(f"wait 120s")
20
+ # sleep(120)
21
+
22
+ def get_free_gpu(threshold=0.95):
23
+ # 获取所有可用的GPU设备
24
+ gpus = GPUtil.getGPUs()
25
+ # 筛选出显存空闲率高于阈值的GPU
26
+ available_gpus = [gpu.id for gpu in gpus if gpu.memoryFree > 80000 and gpu.id not in [4,5]]
27
+ print(f"available_gpus: {available_gpus}")
28
+ return available_gpus
29
+
30
+ def run(commands):
31
+ gpus = get_free_gpu()
32
+ if NEED_REVERSE:
33
+ gpus.reverse()
34
+ # random.shuffle(gpus)
35
+ # print(f"available_gpus: {gpus}")
36
+ while not commands.empty() and len(gpus) >= 2:
37
+ command = commands.get()
38
+ print(f"The following command is about to run:\n{command.format(device=f'{gpus[0]},{gpus[1]}')}")
39
+ subprocess.Popen(command.format(device=f'{gpus[0]},{gpus[1]}'), shell=True, start_new_session=True)
40
+ gpus = gpus[2:]
41
+ sleep(SLEEP_TIME)
42
+ return commands
43
+
44
+ def main():
45
+ global commands, wait_cnt, eppt, end_ckpt, only_rank
46
+ eppt = ckpt
47
+ math_datasets = [
48
+ 'AIME24',
49
+ ]
50
+ # qa_datasets = [
51
+ # 'aime', 'hotpotqa','2wiki', 'musique', "amc", "math500"
52
+ # ]
53
+ # qa_datasets = [
54
+ # 'simpleqa', 'hotpotqa'
55
+ # ]
56
+ # qa_datasets = [
57
+ # 'eval', "realqa"
58
+ # ]
59
+ qa_datasets = [
60
+ 'gaia'
61
+ ]
62
+ ckpt_dir = f"{ckpt}"
63
+
64
+ all_items = os.listdir(ckpt_dir)
65
+ step_list = []
66
+ for item in all_items:
67
+ if item.startswith("checkpoint-"):
68
+ try:
69
+ step = int(item.split("-")[1])
70
+ step_list.append(step)
71
+ except ValueError:
72
+ continue
73
+
74
+ step_list.sort(reverse=True) # 从大到小排序
75
+ # print(f"step_list: {step_list}")
76
+ # # step_list = step_list[:2] # 只取最新的checkpoint
77
+ print(f"step_list: {step_list}")
78
+ step_list = step_list[:4]
79
+ print(f"step_list: {step_list}")
80
+ for step in step_list:
81
+ checkpoint_dir = f"{ckpt}/checkpoint-{step}"
82
+ export_dir = checkpoint_dir
83
+ wait_cnt = 0
84
+ print(f"Waiting for checkpoint ({checkpoint_dir}) to exist...")
85
+
86
+ while not os.path.exists(os.path.join(checkpoint_dir, "special_tokens_map.json")):
87
+ commands = run(commands)
88
+ wait_cnt += 1
89
+ print(f"Already waiting {datetime.timedelta(seconds=SLEEP_TIME*wait_cnt)}.")
90
+ if wait_cnt > (86400 // SLEEP_TIME):
91
+ clear()
92
+ raise Exception("There have been no new checkpoints for no less than a day, and the program is about to automatically exit.")
93
+ # python3 = "/opt/aps/workdir/miniforge3/envs/search_o1/bin/python3"
94
+ python3 = "/opt/aps/workdir/search_o1/bin/python3"
95
+ # /opt/aps/workdir/miniforge3/envs/search_o1/bin/python
96
+ print("The checkpoint exists. Waiting for running...")
97
+ for dataset in qa_datasets:
98
+ base_name = f"{os.path.basename(ckpt)}/{step}/{dataset}"
99
+ dir = "/opt/aps/workdir/sunshuang/deep_search/search_o1"
100
+ # cache_path = f"{dir}/cache_reason_two_model/eval_reason_two_model/qwen-instruct-32B/{base_name}"
101
+ cache_path = f"{dir}/cache_eval_sum_all_webpage"
102
+ output_path = f"{dir}/output/output_eval/{base_name}"
103
+ log_path = f"{dir}/logs/log_eval/{base_name}"
104
+ os.makedirs(cache_path, exist_ok=True)
105
+ os.makedirs(output_path, exist_ok=True)
106
+ os.makedirs(log_path, exist_ok=True)
107
+ contents = os.listdir(output_path)
108
+
109
+ # subset_num = 500
110
+ # if dataset == "math500":
111
+ # subset_num = 500
112
+
113
+
114
+ if not contents:
115
+ command = (
116
+ "export https_proxy=http://127.0.0.1:7890 && "
117
+ "export http_proxy=http://127.0.0.1:7890 && "
118
+ "export all_proxy=socks5://127.0.0.1:7891 && "
119
+ "cd /opt/aps/workdir/sunshuang/deep_search/search_o1 && "
120
+ "CUDA_VISIBLE_DEVICES={device} "
121
+ f"nohup {python3} -u scripts/search_o1_sum_all_webpage.py "
122
+ f"--dataset_name {dataset} "
123
+ f"--cache_dir_base {cache_path} "
124
+ f"--output_dir_base {output_path} "
125
+ "--split test "
126
+ "--max_search_limit 10 "
127
+ "--max_turn 10 "
128
+ "--top_k 10 "
129
+ "--max_doc_len 5000 "
130
+ f"--model_path {checkpoint_dir} "
131
+ "--bing_subscription_key \"cb0d28279a826d7e5cf22d71f683c77ffd4ba27d\" "
132
+ "--bing_endpoint \"https://google.serper.dev/search\" "
133
+ f"> {log_path}/inf.log 2>&1 &"
134
+ )
135
+ print(f"command: {command}")
136
+ commands.put(command)
137
+ else:
138
+ print(f"skip evaluated model: {base_name}")
139
+
140
+ # continue
141
+
142
+ # for dataset in math_datasets:
143
+ # dir = "/opt/aps/workdir/math/evaluation"
144
+ # base_name = f"{os.path.basename(ckpt)}/{step}/{dataset}"
145
+ # output_path = f"{dir}/outputs_maxtoken_20000/{base_name}"
146
+ # log_path = f"{dir}/logs_maxtoken_20000/{base_name}"
147
+ # os.makedirs(output_path, exist_ok=True)
148
+ # os.makedirs(log_path, exist_ok=True)
149
+ # contents = os.listdir(output_path)
150
+ # if not contents:
151
+ # command = (
152
+ # "cd /opt/aps/workdir/math/evaluation && "
153
+ # "CUDA_VISIBLE_DEVICES={device} "
154
+ # f"nohup {python3} -u run_eval_32b.py "
155
+ # f"--data_name {dataset} "
156
+ # f"--target_path {output_path} "
157
+ # f"--model_name_or_path {checkpoint_dir} "
158
+ # "--prompt v4 "
159
+ # "--max_tokens 20000 "
160
+ # "--paralle_size 2 "
161
+ # f"> {log_path}/inf.log 2>&1 &"
162
+ # )
163
+ # commands.put(command)
164
+ # else:
165
+ # print(f"skip evaluated model: {base_name}")
166
+
167
+ def clear():
168
+ print("All checkpoints exist. Wait for runing...")
169
+ global commands
170
+ while not commands.empty():
171
+ # sleep(SLEEP_TIME)
172
+ commands = run(commands)
173
+ print("Wish me good luck!")
174
+
175
+
176
+ if __name__ == '__main__':
177
+ # print(sys.argv[1])
178
+ ckpt = sys.argv[1]
179
+ main()
180
+ clear()
ssh_node.sh ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ ssh lmlabide-87fe1717-6eec-4fb7-94c7-4e7919588b88-worker-0
2
+ ssh lmlabide-87fe1717-6eec-4fb7-94c7-4e7919588b88-worker-1
3
+ ssh lmlabide-87fe1717-6eec-4fb7-94c7-4e7919588b88-worker-2
4
+ ssh lmlabide-87fe1717-6eec-4fb7-94c7-4e7919588b88-worker-3
5
+ ssh lmlabide-87fe1717-6eec-4fb7-94c7-4e7919588b88-worker-4
test.py ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ # import subprocess
3
+
4
+ # def check_gpu_processes_with_nvitop(gpu_id):
5
+ # try:
6
+ # # 调用 nvitop 命令，指定 GPU ID 和单次查询模式
7
+ # result = subprocess.run(
8
+ # ["nvitop", "-o", str(gpu_id), "--once"],
9
+ # stdout=subprocess.PIPE,
10
+ # stderr=subprocess.PIPE,
11
+ # text=True
12
+ # )
13
+
14
+ # # 检查命令是否执行成功
15
+ # if result.returncode != 0:
16
+ # print(f"Error running nvitop: {result.stderr} [[5]]")
17
+ # return True # 如果错误就返回True，即当前GPU不可用
18
+
19
+ # # 解析输出内容
20
+ # has_processes = True
21
+ # len_res = len(result.stdout.splitlines())
22
+
23
+ # no_processes_str = "No running processes found"
24
+ # for line in result.stdout.splitlines():
25
+ # line = line.strip()
26
+ # if no_processes_str in line:
27
+ # print("No running processes found.")
28
+ # has_processes = False
29
+
30
+ # return has_processes
31
+
32
+ # except FileNotFoundError:
33
+ # print("nvitop not found. Install it via 'pip install nvitop' [[4]]")
34
+ # return True # 如果错误就返回True，即当前GPU不可用
35
+
36
+ # # 调用函数，指定 GPU 0
37
+ # print(check_gpu_processes_with_nvitop(1))
38
+
39
+ # 初始化变量 ll
40
+ l1 = [3, [66, 55, 44], (7, 8, 9)]
41
+ l2 = list(l1) # 浅复制了 ll
42
+
43
+ # 输出l1 l2每个元素的id
44
+ print("l1")
45
+ print(id(l1[0]), id(l1[1]), id(l1[2]))
46
+ print("l2")
47
+ print(id(l2[0]), id(l2[1]), id(l2[2]))
48
+
49
+ l1.append(100)
50
+ l1[1].remove(55)
51
+ print("l1")
52
+ print(id(l1[0]), id(l1[1]), id(l1[2]))
53
+ print("l2")
54
+ print(id(l2[0]), id(l2[1]), id(l2[2]))
55
+ l2[1] += [33, 22]
56
+ l2[2] += (10, 11)
57
+ print(id(l1[0]), id(l1[1]), id(l1[2]))
58
+ print(id(l2[0]), id(l2[1]), id(l2[2]))
train_requirements.txt ADDED
@@ -0,0 +1,122 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ numpy==1.26.4
2
+ rouge-chinese==1.0.3
3
+ fire==0.7.0
4
+ transformers==4.46.0
5
+ torch==2.1.0
6
+
7
+ accelerate==0.26.0
8
+ aiofiles==23.2.1
9
+ aiohappyeyeballs==2.4.6
10
+ aiohttp==3.11.12
11
+ aiosignal==1.3.2
12
+ altair==5.5.0
13
+ anyio==4.8.0
14
+ attrs==25.1.0
15
+ certifi==2025.1.31
16
+ charset-normalizer==3.4.1
17
+ click==8.1.8
18
+ contourpy==1.3.1
19
+ cycler==0.12.1
20
+ datasets==2.15.0
21
+ deepspeed==0.14.5
22
+ dill==0.3.7
23
+ einops==0.8.1
24
+ fastapi==0.95.1
25
+ ffmpy==0.5.0
26
+ filelock==3.17.0
27
+
28
+
29
+ fonttools==4.56.0
30
+ frozenlist==1.5.0
31
+ fsspec==2023.10.0
32
+ gputil==1.4.0
33
+ gradio==3.38.0
34
+ gradio-client==1.7.1
35
+ h11==0.14.0
36
+ hjson==3.1.0
37
+ httpcore==1.0.7
38
+ httpx==0.28.1
39
+ huggingface-hub==0.29.1
40
+ idna==3.10
41
+ jieba==0.42.1
42
+ jinja2==3.1.5
43
+ joblib==1.4.2
44
+ jsonschema==4.23.0
45
+ jsonschema-specifications==2024.10.1
46
+ kiwisolver==1.4.8
47
+ linkify-it-py==2.0.3
48
+ markdown-it-py==2.2.0
49
+ markupsafe==2.1.5
50
+ matplotlib==3.10.0
51
+ mdit-py-plugins==0.3.3
52
+ mdurl==0.1.2
53
+ mpmath==1.3.0
54
+ multidict==6.1.0
55
+ multiprocess==0.70.15
56
+ narwhals==1.27.1
57
+ networkx==3.4.2
58
+ ninja==1.11.1.3
59
+ nltk==3.9.1
60
+
61
+ nvidia-cublas-cu12==12.1.3.1
62
+ nvidia-cuda-cupti-cu12==12.1.105
63
+ nvidia-cuda-nvrtc-cu12==12.1.105
64
+ nvidia-cuda-runtime-cu12==12.1.105
65
+ nvidia-cudnn-cu12==8.9.2.26
66
+ nvidia-cufft-cu12==11.0.2.54
67
+ nvidia-curand-cu12==10.3.2.106
68
+ nvidia-cusolver-cu12==11.4.5.107
69
+ nvidia-cusparse-cu12==12.1.0.106
70
+ nvidia-ml-py==12.570.86
71
+ nvidia-nccl-cu12==2.18.1
72
+ nvidia-nvjitlink-cu12==12.8.61
73
+ nvidia-nvtx-cu12==12.1.105
74
+ orjson==3.10.15
75
+ packaging==24.2
76
+ pandas==2.2.3
77
+ peft==0.6.1
78
+ pillow==10.4.0
79
+ propcache==0.3.0
80
+ protobuf==5.29.3
81
+ psutil==7.0.0
82
+ py-cpuinfo==9.0.0
83
+ pyarrow==19.0.1
84
+ pyarrow-hotfix==0.6
85
+ pydantic==1.10.11
86
+ pydub==0.25.1
87
+ pyparsing==3.2.1
88
+ python-dateutil==2.9.0.post0
89
+ python-multipart==0.0.20
90
+ pytz==2025.1
91
+ pyyaml==6.0.2
92
+ referencing==0.36.2
93
+ regex==2024.11.6
94
+ requests==2.32.3
95
+
96
+ rpds-py==0.23.1
97
+ safetensors==0.5.2
98
+ scipy==1.15.2
99
+ semantic-version==2.10.0
100
+ sentencepiece==0.2.0
101
+ six==1.17.0
102
+ sniffio==1.3.1
103
+ sse-starlette==2.1.3
104
+ starlette==0.26.1
105
+ sympy==1.13.3
106
+ termcolor==2.5.0
107
+ tiktoken==0.9.0
108
+ tokenizers==0.20.3
109
+
110
+ tqdm==4.67.1
111
+
112
+ triton==2.1.0
113
+ typing-extensions==4.12.2
114
+ tzdata==2025.1
115
+ uc-micro-py==1.0.3
116
+ urllib3==2.3.0
117
+ uvicorn==0.34.0
118
+ websockets==11.0.3
119
+ xxhash==3.5.0
120
+ yarl==1.18.3
121
+
122
+ flash-attn==2.7.1.post1