Skip to content

Switch to Qwen3-Embedding-4B and setup sklearn.model_selection.GroupKFold to sold character mapping, overfitting problem #7

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
rjurney wants to merge 2 commits into
base: main
Choose a base branch
from
Open

Switch to Qwen3-Embedding-4B and setup sklearn.model_selection.GroupKFold to sold character mapping, overfitting problem #7

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
56c1acf
Working my dependencies towards installing flash-attn for Qwen3-Embed…
rjurney Oct 1, 2025
39b17a5
Use Qwen3 across the board and install flash-attn using pip
rjurney Oct 1, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
7 changes: 5 additions & 2 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -202,6 +202,9 @@ curl -sSL https://install.python-poetry.org | python3 -
```bash
# Install dependencies
poetry install

# Install flash-attn for Qwen3-Embedding-4B, it does not support PEP-517 builds, so no Poetry
pip install flash-attn --no-build-isolation
```

### Optional: Weights and Biases for Experiment Tracking
Expand Down Expand Up @@ -286,10 +289,10 @@ wandb login
# Using FP16 mixed precision for faster training
# Using a sample fraction of 0.5 to use half the dataset
nohup eridu train \
--model "intfloat/multilingual-e5-base" \
--model "Qwen/Qwen3-Embedding-4B" \
--use-gpu \
--fp16 \
--batch-size 32 \
--batch-size 128 \
--epochs 1 \
--patience 1 \
--weight-decay 0.01 \
Expand Down
32 changes: 31 additions & 1 deletion poetry.lock
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

2 changes: 2 additions & 0 deletions pyproject.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -88,6 +88,8 @@ cuml-cu12 = { version = "^25.6.0", source = "rapids" }
cleanco = "^2.3"
stop-words = "^2018.7.23"
langdetect = "^1.0.9"
packaging = "^25.0"
ninja = "^1.13.0"

[tool.poetry.group.dev.dependencies]
black = { version = "^25.1.0", extras = ["jupyter"] }
Expand Down