Training Medusa with QLoRA + FSDP
active 2024-01-14 → 2026-06-25 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 894 days from 2024-01-14 to 2026-06-25. Pushes: 260 total, peak 12 in a day. Pull requests: 64 total, peak 9 in a day. Issues: 46 total, peak 3 in a day. Comments: 133 total, peak 16 in a day. Stars: 1,454 total, peak 282 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 |
|---|---|---|---|---|
| KeremTurgutlu | 202 | 184 | 8 | 6 |
| warner-benjamin | 63 | 33 | 17 | 10 |
| johnowhitaker | 62 | 29 | 20 | 10 |
| hsb1995 | 22 | 0 | 0 | 20 |
| austinvhuang | 16 | 4 | 3 | 5 |
| geronimi73 | 15 | 0 | 3 | 11 |
| sanipanwala | 13 | 0 | 0 | 13 |
| iseesaw | 10 | 0 | 0 | 9 |
| griff4692 | 9 | 9 | 0 | 0 |
| jph00 | 7 | 1 | 0 | 4 |
| rationalism | 6 | 0 | 0 | 3 |
| bilalghanem | 6 | 0 | 0 | 4 |
| jeromeku | 6 | 0 | 1 | 4 |
| Xynonners | 5 | 0 | 0 | 4 |
| ehartford | 5 | 0 | 0 | 3 |
| chwenjun225 | 5 | 0 | 2 | 1 |
| Pugio | 4 | 0 | 0 | 3 |
| catid | 4 | 0 | 0 | 2 |
| mrgohlke | 3 | 0 | 0 | 1 |
| deepankarsharma | 3 | 0 | 1 | 1 |
Recent activity
Latest issues, pull requests and releases
- Issue comment#76Ruiwen5052025-05-19 14:18Bitsandbytes error while run train.py
- Issue comment#76matthewdouglas2025-05-13 19:57Bitsandbytes error while run train.py
- Issue comment#76Ruiwen5052025-05-12 14:31Bitsandbytes error while run train.py
- Issue comment#76matthewdouglas2025-05-10 00:42Bitsandbytes error while run train.py
- Issue comment#76geronimi732025-05-08 20:05Bitsandbytes error while run train.py
- Issue#76Ruiwen5052025-05-08 14:52Bitsandbytes error while run train.py
- Issue#75yzhang1232025-03-25 15:32NotImplementedError: c10d::broadcast_: at
- Issue#74alon-123452025-01-29 16:11Training using FSDP, qLoRa on multinode
- Pull request#62geronimi732025-01-09 10:22
- Pull request#71chwenjun2252024-12-27 06:15
- Issue comment#73ghsama2024-12-13 11:06Out of memory while ConvertingTheStateDict - How to split across GPUs?
- Issue comment#45mrgohlke2024-11-06 19:14nan when the input length is large
- Issue#73mrgohlke2024-10-19 17:40Out of memory while ConvertingTheStateDict - How to split across GPUs?
- Issue#73mrgohlke2024-10-19 17:20Out of memory while ConvertingTheStateDict - How to split across GPUs?
- Pull request#72chrismrutherford2024-09-13 09:57
- Pull request#71chwenjun2252024-08-31 16:37
- Issue#70chwenjun2252024-08-31 16:14`Converting the State Dict.ipynb` - Runtine error because Unexpected keys
- Issue comment#70chwenjun2252024-08-31 16:13`Converting the State Dict.ipynb` - Runtine error because Unexpected keys
- Issue#70chwenjun2252024-08-31 15:23`Converting the State Dict.ipynb` - Runtine error because Unexpected keys
- Issue#69asmith262024-08-31 15:05How to fine-tune a Vision Language Model (VLM)?
- Issue#68BenjaminBossan2024-08-15 10:00DoRA training not taking dropout or alpha into account
- Issue comment#60williambarberjr2024-07-13 22:59Request for Scripts to Merge QDoRA Adapters with Base Model for vLLM Inference
- Issue comment#60lochuynh14122024-06-18 18:41Request for Scripts to Merge QDoRA Adapters with Base Model for vLLM Inference
- Pull request#67austinvhuang2024-06-11 03:37
- Issue comment#66jph002024-06-09 08:20Add profiling to train.py
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 1,454 stars here means stars gained during the window, not the repo's star count.