Card: upload status, weights_only=True
Browse files
README.md
CHANGED
|
@@ -17,11 +17,12 @@ continued pretraining from the final weights of [Slayer1B 1.08](https://huggingf
|
|
| 17 |
optimizer state. These are PyTorch files written by the trainer, for resuming or studying training — not a
|
| 18 |
Transformers/safetensors package and not a finished model.
|
| 19 |
|
| 20 |
-
**Status: training in progress
|
| 21 |
-
to update 35,000 and one full-state checkpoint at update 35,000.
|
| 22 |
-
|
| 23 |
-
|
| 24 |
-
|
|
|
|
| 25 |
|
| 26 |
| | |
|
| 27 |
|---|---|
|
|
@@ -41,7 +42,7 @@ object id (sha256) and the size match.
|
|
| 41 |
| `wagi_<update>.pt` | model weights (float32), `config`, `step` | 4.3 GB |
|
| 42 |
| `ckpt_35000.pt` | model weights, optimizer state (Muon and AdamW), data-queue position, `config`, `step` | 8.9 GB |
|
| 43 |
|
| 44 |
-
`wagi` is Polish for “weights”. Load with `torch.load(path, map_location="cpu")`; the model definition is in
|
| 45 |
`train_gpt_ref_r23.py` in the [1.08 repository](https://huggingface.co/SlayerLab/Slayer1B-1.08).
|
| 46 |
|
| 47 |
## Tokens per checkpoint
|
|
|
|
| 17 |
optimizer state. These are PyTorch files written by the trainer, for resuming or studying training — not a
|
| 18 |
Transformers/safetensors package and not a finished model.
|
| 19 |
|
| 20 |
+
**Status: training in progress; files are being uploaded (12 November 2026).** The tables below list the target set: every
|
| 21 |
+
weights-only checkpoint saved at a 5,000-update boundary up to update 35,000 and one full-state checkpoint at update 35,000.
|
| 22 |
+
Files appear one at a time; a file missing from the file list is not uploaded yet. Later checkpoints will be added as the
|
| 23 |
+
run continues. The training curve and evaluations are on [Track](https://track.fabryka.ai/run/71cf98ec-7c60-54a8-ac47-0ea7a23f573a).
|
| 24 |
+
Each uploaded file is compared after upload with the checksum written when it was copied from the training machine: the LFS
|
| 25 |
+
object id (sha256) and the size must match.
|
| 26 |
|
| 27 |
| | |
|
| 28 |
|---|---|
|
|
|
|
| 42 |
| `wagi_<update>.pt` | model weights (float32), `config`, `step` | 4.3 GB |
|
| 43 |
| `ckpt_35000.pt` | model weights, optimizer state (Muon and AdamW), data-queue position, `config`, `step` | 8.9 GB |
|
| 44 |
|
| 45 |
+
`wagi` is Polish for “weights”. Load with `torch.load(path, map_location="cpu", weights_only=True)`; the model definition is in
|
| 46 |
`train_gpt_ref_r23.py` in the [1.08 repository](https://huggingface.co/SlayerLab/Slayer1B-1.08).
|
| 47 |
|
| 48 |
## Tokens per checkpoint
|