大模型剪枝 A simple and effective LLM pruning approach.
active 2023-08-15 → 2026-04-15 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 975 days from 2023-08-15 to 2026-04-15. Pushes: 22 total, peak 8 in a day. Pull requests: 21 total, peak 2 in a day. Issues: 70 total, peak 7 in a day. Comments: 116 total, peak 14 in a day. Stars: 507 total, peak 6 in a day.
- Pushes
- Pull requests
- Issues
- Comments
- Stars
Stars, PRs, issues and forks are under-captured in the later part of this window. GH Archive progressively stopped capturing non-push events during 2026 — −95% or worse by the end of the window. Every series here except Pushes fades for that reason, so a decline above reflects the archive, not this repository. Pushes stay reliable throughout, so read them, and the contributor counts derived from them, as the real signal. Data health has the measurements.
Top contributors
Pushes, PRs, issues, reviews and comments — stars and forks excluded, so this is contribution rather than popularity
| Contributor | Contributions | Pushes | PRs | Comments |
|---|---|---|---|---|
| Eric-mingjie | 60 | 22 | 0 | 31 |
| junzhang-zj | 12 | 0 | 0 | 11 |
| NamburiSrinath | 11 | 0 | 0 | 7 |
| taratt | 7 | 0 | 0 | 5 |
| hainingfang | 6 | 0 | 0 | 2 |
| pprp | 4 | 0 | 0 | 2 |
| irugina1 | 4 | 0 | 4 | 0 |
| simlaharma | 4 | 0 | 0 | 2 |
| Logan-007L | 4 | 0 | 0 | 4 |
| Qian2333 | 4 | 0 | 0 | 2 |
| egeor | 3 | 0 | 0 | 2 |
| JiwenJ | 3 | 0 | 0 | 1 |
| dat-adi | 3 | 0 | 0 | 3 |
| lofty11 | 3 | 0 | 0 | 1 |
| Shinning-Zhou | 3 | 0 | 0 | 2 |
| TagoreZhao | 3 | 0 | 2 | 1 |
| kast424 | 3 | 0 | 0 | 2 |
| ronimalka | 3 | 0 | 2 | 1 |
| frankinwi | 3 | 0 | 0 | 1 |
| virentakia | 2 | 0 | 0 | 1 |
Recent activity
Latest issues, pull requests and releases
- Issue comment#82hide-on-bush-lalala2025-10-28 06:44Issue with Wanda pruning on Llama 3.2-1b - Stuck at "loading calibration data
- Issue comment#82SuchitGupte2025-10-23 17:19Issue with Wanda pruning on Llama 3.2-1b - Stuck at "loading calibration data
- Issue#85hxu1052025-07-31 17:45The role of calibration data
- Issue comment#76Pengxin-Guo2025-07-25 15:14what's the version of lm_eval?
- Pull request#84jonny-vr2025-07-08 10:17
- Pull request#84jonny-vr2025-07-08 10:16
- Issue comment#80LyoAI2025-06-27 08:07Differences in Perplexity from the reported values
- Issue comment#82hide-on-bush-lalala2025-06-18 12:10Issue with Wanda pruning on Llama 3.2-1b - Stuck at "loading calibration data
- Issue#83BenWu112025-04-27 15:07Unexpected Sparsity Drop After LoRA Finetuning with Provided Scripts
- Issue comment#44dat-adi2025-04-27 06:14gpu memory size recommended for pruning the llama2-7b-chat-hf model
- Issue#82xay20012025-04-23 07:20Issue with Wanda pruning on Llama 3.2-1b - Stuck at "loading calibration data
- Issue comment#44dat-adi2025-04-20 16:28gpu memory size recommended for pruning the llama2-7b-chat-hf model
- Issue comment#44dat-adi2025-04-20 16:21gpu memory size recommended for pruning the llama2-7b-chat-hf model
- Issue comment#62xiejingcheng2025-04-10 10:33AttributeError: module 'lm_eval.tasks' has no attribute 'ALL_TASKS'
- Issue#81davidray2222025-03-20 19:31the question about LLM inference performance
- Issue comment#80LawrenceRLiu2025-03-15 00:32Slight Differences in Perplexity for Llama 2 7b
- Issue#80LawrenceRLiu2025-03-15 00:14Slight Differences in Perplexity for Llama 2 7b
- Issue comment#44junzhang-zj2025-02-27 13:23gpu memory size recommended for pruning the llama2-7b-chat-hf model
- Issue#79hyleepp2025-02-27 07:35Does the padding tokens in a batch will be calculated?
- Issue comment#68junzhang-zj2025-02-22 14:45Memory and time requirements for Mistral-7B
- Issue comment#44shubham-maindola2025-02-19 17:07gpu memory size recommended for pruning the llama2-7b-chat-hf model
- Pull request#78diaconuccalin2025-02-13 16:36
- Pull request#78diaconuccalin2025-02-13 16:36
- Issue comment#68NamburiSrinath2025-01-15 03:23Memory and time requirements for Mistral-7B
- Issue#68NamburiSrinath2025-01-15 03:23Memory and time requirements for Mistral-7B
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 507 stars here means stars gained during the window, not the repo's star count.