issues
search
Kipok
/
NeMo-Skills
A pipeline to improve skills of large language models
Apache License 2.0
147
stars
33
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Added new benchmark AMC23, corrected an error in AIME-2024, and removed leading zeros from AIME-2024 for improved formatting.
#118
wedu-nvidia
opened
3 hours ago
0
Added new benchmark AMC23, corrected an error in AIME-2024, and removed leading zeros from AIME-2024 for improved formatting
#117
wedu-nvidia
closed
3 hours ago
0
Added new benchmark AMC23, corrected an error in AIME-2024, and removed leading zeros from AIME-2024 for improved formatting
#116
wedu-nvidia
closed
5 hours ago
0
Added new benchmark AMC23, corrected an error in AIME-2024, and removed leading zeros from AIME-2024 for improved formatting
#115
wedu-nvidia
closed
5 hours ago
0
Migrating #88 (decontamination pipeline)
#114
Kipok
closed
1 day ago
0
Integrate #107 with fixes for examples type logic
#113
Kipok
closed
2 days ago
0
Add uncommited changes check
#112
Kipok
closed
2 days ago
0
Refactor of #105
#111
Kipok
closed
2 days ago
0
Fixes for checkpoint averaging with steps
#110
Kipok
closed
3 days ago
0
Upgrade nemo-run to store logs close to results
#109
Kipok
closed
3 days ago
0
Integrating #89 + upgrade to sft docker + some renaming
#108
Kipok
closed
3 days ago
0
Updated prompts
#107
i-vainn
closed
2 days ago
1
Update SDP pipeline
#106
i-vainn
opened
4 days ago
0
Summarize results from cluster
#105
i-vainn
closed
2 days ago
1
Added uncommitted code check
#104
i-vainn
closed
2 days ago
1
Add config folder argument
#103
i-vainn
closed
5 days ago
0
Added dataset groups to prepare.py
#102
i-vainn
closed
5 days ago
0
Checkpoint conversion + check_if_mounted logic
#101
Kipok
closed
5 days ago
0
Initial NeMo-Run integration
#100
Kipok
closed
5 days ago
0
Fix regex escape for parsing of end tokens
#99
titu1994
closed
2 weeks ago
0
Modify sandbox code execution to capture stderr
#98
aleksficek
closed
1 week ago
3
update paths in the README file
#97
dgtm777
closed
3 weeks ago
0
Fix tests for NeMo Inspector
#96
dgtm777
closed
3 weeks ago
0
Remove speicifc commit from SDP installation
#95
Kipok
closed
3 weeks ago
0
fix of imports
#94
dgtm777
closed
3 weeks ago
0
Filtering long solutions out
#93
shtoshni
closed
3 weeks ago
0
Add support to change vllm max num sequences processed
#92
titu1994
closed
3 weeks ago
0
Rename the tool
#91
dgtm777
closed
3 weeks ago
0
Prompt refactoring
#90
Kipok
closed
3 weeks ago
0
WIP - Llama code interpreter and improved LLM as a judge
#89
Kipok
closed
3 days ago
1
push decontaminator code
#88
wedu-nvidia
closed
1 day ago
1
Upgrade trtllm to dev2024080600
#87
Kipok
closed
1 month ago
0
Fix fill output args
#86
yinglin-nv
closed
1 month ago
1
change default tokens_to_generate to 2k
#85
bmwshop
closed
1 month ago
0
shtoshni/code sft data preparation
#84
shtoshni
closed
1 month ago
2
Reverting optimization changes for stable training runs
#83
shtoshni
closed
1 month ago
0
Optimizing SFT recipe
#82
Kipok
closed
1 month ago
1
Multiple fixes
#81
Kipok
closed
1 month ago
0
Shtoshni/llama3.1 ft
#80
shtoshni
closed
1 month ago
1
Generate batch support
#79
Alcray
opened
1 month ago
0
Remove `tokenizer_type` when loading a tokenizer using AutoTokenizer
#78
yinglin-nv
closed
1 month ago
0
Upgrade trtllm to support llama 3.1
#77
Kipok
closed
1 month ago
0
Update vLLM
#76
titu1994
closed
1 month ago
1
Berkeley Function Calling Leaderboard (BFCL) Integration
#75
shtoshni
opened
1 month ago
2
Adding llama3 prompts + llm-as-a-judge eval
#74
Kipok
closed
1 month ago
0
Igitman/test
#73
Kipok
closed
1 month ago
0
DPO (RPO) scripts integration
#72
Kipok
closed
1 month ago
0
Proper metrics for ifeval
#71
Kipok
closed
1 month ago
0
GPU tests in CI
#70
Kipok
closed
1 month ago
1
Add editable fields to the Data Explorer
#69
dgtm777
closed
1 month ago
0
Next