Downloads · 30 days
0
finsearch-x-qwen3-4b-mlx
xingyujeff/finsearch-x-qwen3-4b-mlx
finsearch-x-qwen3-4b-mlx is a machine learning model from xingyujeff. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
Status: PRETRAINING / feasibility evidence only. No production model or scaling result is claimed.
Access
Public
Updated Sep 14, 2026
Repo size
—
Likes
1
Public
Repo makeup
Click a slice to open those files.
.json66.2 KB · 48%
At a glance
- Access
- Public
- Created
- Sep 14, 2026
- Updated
- Sep 14, 2026
- SHA
- 71860c99
Model card
From the Hugging Face model README
FinSearch-X Qwen3-4B MLX experiment (draft)
Status: PRE_TRAINING / feasibility evidence only. No production model or scaling result is claimed.
- Backend: MLX; quantized base:
mlx-community/Qwen3-4B-Instruct-2507-4bit. - Base revision:
50d427756c6b1b2fe0c0a10f67fbda1fc8e82c1b. - Parent:
Qwen/Qwen3-4B-Instruct-2507atcdbee75f17c01a7cc42f958dc650907174af0554. - Task: evidence-grounded claim selection and answerability.
- Dataset:
FINSEARCH_X_SFT_DATASET_v1_RC3; training release remains false. - Planned arms: BASE, SFT-500, SFT-1000, SFT-1600, with nested train subsets.
Phase 2.3A established local model load, base generation, data projection, assistant-only loss masking, LoRA smoke training, adapter reload, a memory ladder, and LR calibration. Formal scaling is not completed and no future result is filled in here.
- Created
- Sep 14, 2026
- Updated
- Sep 14, 2026