Downloads · 30 days
0
destinefut/nanochat-d18-sft
nanochat-d18-sft is a machine learning model from destinefut. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for generic. The card lists the license as mit.
The model configuration (derived from meta002849.json) is as follows:
Downloads · 30 days
0
Access
Public
Updated Mar 7, 2026
Repo size
3 GB
Likes
0
Public
Click a slice to open those files.
.pt3 GB · 100%
From the Hugging Face model README
The model configuration (derived from meta_002849.json) is as follows:
| Parameter | Value |
|---|---|
| Layers | 18 |
| Embedding Dim | 1152 |
| Heads (Q/KV) | 9 / 9 (GQA) |
| Vocab Size | 65,536 |
| Max Seq Len | 2048 |
| Window Pattern | SSSL |
| Metric | BASE | SFT | RL |
|---|---|---|---|
| CORE | 0.2500 | - | - |
| ARC-Challenge | - | 0.3942 | - |
| ARC-Easy | - | 0.4722 | - |
| GSM8K | - | 0.0387 | - |
| HumanEval | - | 0.0915 | - |
| MMLU | - | 0.3418 | - |
| ChatCORE | - | 0.2882 | - |
This model uses a custom architecture and cannot be loaded directly via the standard transformers library. Please use the source code from the official GitHub repository.
git clone https://github.com/DestineG/nanochat.git
cd nanochat
git checkout v1.0