Automation foundation model for tiny devices: 2-bit, 8-29 MB, tool calls, ASR, structured extraction and embeddings on phones, wearables, smart homes, robots, cars and microcontrollers.
docs: document finetune CLI options (#93)
WeiHaoxuan25ef4ac48231
diff --git a/README.md b/README.mdindex c032aac..f10209b 100644--- a/README.md+++ b/README.md@@ -114,7 +114,7 @@ needle finetune data.jsonl --epochs 10needle finetune data.jsonl --epochs 10 --generate 300 --lora-rank 16 --lora-alpha 32```-Key options: `--epochs` (default 3), `--lora-rank` (16), `--lora-alpha` (32), `--lr` (1e-4), `--batch-size` (16), `--max-len` (1024), `--val-split` (0.1), `--checkpoint <base.pkl>`, `--out <adapter.pkl>`. The adapter is written to `checkpoints/needle_lora.pkl`. A validation loss prints each epoch from the held out split.+Key options: `--epochs` (default 3), `--lora-rank` (16), `--lora-alpha` (32), `--lr` (1e-4), `--batch-size` (16), `--max-len` (1024), `--val-split` (0.1), `--checkpoint <base.pkl>`, `--checkpoint-dir <dir>` (default `checkpoints`), `--out <adapter.pkl>`, `--generate <n>`, `--model <id>` (default `deepseek/deepseek-v4-flash`), and `--workers <n>` (default 8). `--generate` uses the configured OpenRouter endpoint to synthesize extra examples before training. The adapter is written to `checkpoints/needle_lora.pkl` by default. A validation loss prints each epoch from the held out split.Training is plain JAX and runs on any accelerator jax supports. On an NVIDIA machine install the CUDA build and the same command trains on the GPU:diff --git a/llms.txt b/llms.txtindex 9a0b6b3..a9def39 100644--- a/llms.txt+++ b/llms.txt@@ -173,7 +173,7 @@ agent = needle.Needle(weights="my.cact", tools=[...])agent.run("...")```-Options: `finetune --lora-rank 16 --lora-alpha 32 --lr 1e-4 --batch-size 16 --max-len 1024 --checkpoint <base.pkl> --out <adapter.pkl>`; `build --bits 2|4 --upload` (with `NEEDLE_HF_REPO=<you>/<model>`).+Options: `finetune --lora-rank 16 --lora-alpha 32 --lr 1e-4 --batch-size 16 --max-len 1024 --val-split 0.1 --checkpoint <base.pkl> --checkpoint-dir <dir> --out <adapter.pkl> --generate <n> --model <id> --workers <n>`; `--model` defaults to `deepseek/deepseek-v4-flash`, `--workers` defaults to `8`, and `--checkpoint-dir` defaults to `checkpoints`. `--generate` synthesizes extra examples through the configured OpenRouter endpoint before training. `build --bits 2|4 --upload` (with `NEEDLE_HF_REPO=<you>/<model>`).## Playground (browser)