[ACL 2025 Main] EfficientQAT: Efficient Quantization-Aware Training for Large Language Models
active 2024-07-28 → 2026-04-10 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 622 days from 2024-07-28 to 2026-04-10. Pushes: 3 total, peak 1 in a day. Pull requests: 1 total, peak 1 in a day. Issues: 30 total, peak 5 in a day. Comments: 36 total, peak 6 in a day. Stars: 146 total, peak 17 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 |
|---|---|---|---|---|
| ChenMnZ | 37 | 3 | 1 | 20 |
| mxjmtxrm | 5 | 0 | 0 | 2 |
| laomao0 | 4 | 0 | 0 | 2 |
| w32zhong | 3 | 0 | 0 | 2 |
| yancaoweidaode | 3 | 0 | 0 | 2 |
| gdsaikrishna | 3 | 0 | 0 | 2 |
| QB-Chen | 2 | 0 | 0 | 0 |
| HXuan-Wang | 2 | 0 | 0 | 1 |
| kaleid-liner | 2 | 0 | 0 | 2 |
| sihouzi21c | 2 | 0 | 0 | 1 |
| Niko-zyf | 1 | 0 | 0 | 0 |
| LiMa-cas | 1 | 0 | 0 | 0 |
| tayton42 | 1 | 0 | 0 | 1 |
| RYY177 | 1 | 0 | 0 | 1 |
| snps-tonatiuh | 1 | 0 | 0 | 0 |
| LiuSiQi-TJ | 1 | 0 | 0 | 0 |
| bimalmagar10 | 1 | 0 | 0 | 0 |
Recent activity
Latest issues, pull requests and releases
- Issue#34bimalmagar102026-01-19 18:50Block‑AP Training Performance on H100
- Issue comment#21HXuan-Wang2026-01-04 12:40Reproduce Llama3 8B Instruct results
- Issue#33HXuan-Wang2025-12-24 07:27About training setting of Llama3-8B
- Issue comment#21RYY1772025-08-08 08:37Reproduce Llama3 8B Instruct results
- Issue comment#25gdsaikrishna2025-07-07 06:14Llava 1.6 Instruction Finetuning
- Issue comment#29ChenMnZ2025-05-29 01:23How are the quantization parameters of block_ap passed to the e2e-qp stage?
- Issue comment#29tayton422025-05-27 03:03How are the quantization parameters of block_ap passed to the e2e-qp stage?
- Issue comment#29ChenMnZ2025-05-26 09:39How are the quantization parameters of block_ap passed to the e2e-qp stage?
- Issue comment#28ChenMnZ2025-05-21 01:42Release artifacts (models, dataset) on Hugging Face
- Issue comment#26ChenMnZ2024-12-08 10:17其它模型是否支持？
- Issue#24ChenMnZ2024-11-14 00:55Unable to reproduce the result in paper for 4 bit wikitext perplexity
- Issue comment#24ChenMnZ2024-11-01 01:48Unable to reproduce the result in paper for 4 bit wikitext perplexity
- Issue comment#23sihouzi21c2024-10-27 15:19Is it possible to run e2e-qp process on a single 4090?
- Issue#23sihouzi21c2024-10-27 15:19Is it possible to run e2e-qp process on a single 4090?
- Issue comment#23ChenMnZ2024-10-26 01:13Is it possible to run e2e-qp process on a single 4090?
- Issue#17ChenMnZ2024-09-17 06:03Can the results of the quantification process be saved?
- Issue#20ChenMnZ2024-09-17 06:01RuntimeError: Triton Error [CUDA]: device kernel image is invalid
- Issue#22ChenMnZ2024-09-17 06:01any experiments on qwen2-7b-instruct?
- Issue#21gdsaikrishna2024-09-11 06:32Reproduce Llama3 8B Instruct results
- Issue comment#20gdsaikrishna2024-09-11 06:27RuntimeError: Triton Error [CUDA]: device kernel image is invalid
- Issue#20Niko-zyf2024-09-11 03:18RuntimeError: Triton Error [CUDA]: device kernel image is invalid
- Issue comment#9w32zhong2024-08-30 20:54Quip# inference speed in your paper is incorrect?
- Issue comment#18w32zhong2024-08-30 20:50why model output is too much slow?
- Issue#19w32zhong2024-08-30 20:49Is 7B llama speed expected to be slow?
- Issue comment#17ChenMnZ2024-08-21 07:50Can the results of the quantification process be saved?
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 146 stars here means stars gained during the window, not the repo's star count.