Instructions to use PRATYUSH-BHARDWAJ/Cortex_A_0.5 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use PRATYUSH-BHARDWAJ/Cortex_A_0.5 with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("PRATYUSH-BHARDWAJ/Cortex_A_0.5", device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- Unsloth Desktop
metrics @ step 16
Browse files- runs/metrics.jsonl +2 -0
runs/metrics.jsonl
CHANGED
|
@@ -1 +1,3 @@
|
|
| 1 |
{"type": "train", "step": 1, "epoch": 0.0174, "loss": 2.5696372985839844, "train_ppl": 13.061086307224265, "grad_norm": NaN, "lr": 0.0, "elapsed_h": 0.012, "budget_h": 0.5}
|
|
|
|
|
|
|
|
|
| 1 |
{"type": "train", "step": 1, "epoch": 0.0174, "loss": 2.5696372985839844, "train_ppl": 13.061086307224265, "grad_norm": NaN, "lr": 0.0, "elapsed_h": 0.012, "budget_h": 0.5}
|
| 2 |
+
{"type": "train", "step": 10, "epoch": 0.1739, "loss": 3.1206914401716657, "train_ppl": 22.662043674549682, "mtp_loss": null, "grad_norm": NaN, "lr": 7.5e-05, "tokens_per_sec": 184.1, "gpu_mem_gb": 6.5, "elapsed_h": 0.041, "budget_h": 0.5}
|
| 3 |
+
{"type": "train", "step": 10, "epoch": 0.1739, "loss": 2.5477417839898004, "train_ppl": 12.778215206455922, "grad_norm": 28.1390438079834, "lr": 0.0, "elapsed_h": 0.041, "budget_h": 0.5}
|