This project aims to build a small language model for few-shot classification tasks. This includes all of the infrastracture needed to train it, run predictions and compare methodologies. Overall I'm interested in researching how far can we push the performance of small language models.
active 2024-04-05 → 2025-04-01 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 362 days from 2024-04-05 to 2025-04-01. Pushes: 32 total, peak 6 in a day. Pull requests: 0 total, peak 0 in a day. Issues: 22 total, peak 9 in a day. Comments: 1 total, peak 1 in a day. Stars: 9 total, peak 1 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 |
|---|---|---|---|---|
| SerjSmor | 55 | 32 | 0 | 1 |
Recent activity
Latest issues, pull requests and releases
- Issue#16SerjSmor2024-04-28 19:23Add Atis notebooks with hashes
- Issue#15SerjSmor2024-04-28 17:30Add the rest of the clinc_oos dataset
- Issue#10SerjSmor2024-04-28 15:08Add a subsample of clinc_oos dataset
- Issue#4SerjSmor2024-04-28 15:07Add Setfit zero-shot classification on Atis dataset
- Issue#14SerjSmor2024-04-20 21:02OOM exceptions during training
- Issue#14SerjSmor2024-04-19 03:23OOM exceptions during training
- Issue#13SerjSmor2024-04-13 06:43Test training small-T5
- Issue#12SerjSmor2024-04-13 06:10Create an unseen dataset for testing
- Issue#11SerjSmor2024-04-13 06:09Add a subsample of yahoo answers topic classification
- Issue#10SerjSmor2024-04-13 06:08Add a subsample of clinc_oos dataset
- Issue#9SerjSmor2024-04-13 06:04Add a sampled Atis dataset to improve overall performance
- Issue#9SerjSmor2024-04-13 06:04Add a sampled Atis dataset to improve overall performance
- Issue#8SerjSmor2024-04-07 07:57Add CoT to the current train regime
- Issue#7SerjSmor2024-04-07 06:40Add a mode of suggesting the name of the class without any prompt options
- Issue#6SerjSmor2024-04-07 06:36Perform text processing in order to improve zero shot performance
- Issue#5SerjSmor2024-04-07 06:18Add GPT3.5 zero shot classification results on Atis dataset
- Issue comment#3SerjSmor2024-04-07 06:15Improve zero shot by running several times and averaging results
- Issue#4SerjSmor2024-04-07 06:06Add Setfit zero-shot classification on Atis dataset
- Issue#3SerjSmor2024-04-07 05:44Improve zero shot by running several times and averaging results
- Issue#1SerjSmor2024-04-07 05:36Add Bart MNLI zero shot classification on Atis dataset
- Issue#2SerjSmor2024-04-07 05:36Add Flan T5 Base zero shot classification on Atis dataset
- Issue#2SerjSmor2024-04-07 03:43Add Flan T5 Base zero shot classification on Atis dataset
- Issue#1SerjSmor2024-04-05 16:46Add Bart MNLI zero shot classification on Atis dataset
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 9 stars here means stars gained during the window, not the repo's star count.