Ship model from checkpoint-3650 via Forge
Browse files- README.md +20 -0
- config.json +25 -0
- forge_metadata.json +10 -0
- generation_config.json +9 -0
- model.safetensors +3 -0
- training_args.bin +3 -0
README.md
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: other
|
| 3 |
+
library_name: transformers
|
| 4 |
+
tags:
|
| 5 |
+
- forge
|
| 6 |
+
- nomad-training
|
| 7 |
+
---
|
| 8 |
+
|
| 9 |
+
# ember-275m
|
| 10 |
+
|
| 11 |
+
This model was trained using [Forge](https://github.com/your-repo/forge).
|
| 12 |
+
|
| 13 |
+
## Training Details
|
| 14 |
+
- **Final Step**: 3650
|
| 15 |
+
- **Final Loss**: 1.6869081497192382
|
| 16 |
+
- **Hardware**: modal
|
| 17 |
+
- **Orchestration**: Nomad Training via Forge
|
| 18 |
+
|
| 19 |
+
## Checkpoint Source
|
| 20 |
+
Exported from atomic checkpoint `3650` on the `checkpoints` branch.
|
config.json
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"EmberForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"bos_token_id": 1,
|
| 6 |
+
"dropout": 0.0,
|
| 7 |
+
"dtype": "float32",
|
| 8 |
+
"eos_token_id": 2,
|
| 9 |
+
"head_dim": 64,
|
| 10 |
+
"hidden_size": 1024,
|
| 11 |
+
"intermediate_size": 2730,
|
| 12 |
+
"max_position_embeddings": 2048,
|
| 13 |
+
"model_type": "ember",
|
| 14 |
+
"num_attention_heads": 16,
|
| 15 |
+
"num_hidden_layers": 18,
|
| 16 |
+
"num_key_value_heads": 8,
|
| 17 |
+
"pad_token_id": 3,
|
| 18 |
+
"rms_norm_eps": 1e-06,
|
| 19 |
+
"rope_theta": 10000.0,
|
| 20 |
+
"tie_word_embeddings": true,
|
| 21 |
+
"transformers_version": "5.9.0",
|
| 22 |
+
"use_bias": false,
|
| 23 |
+
"use_cache": false,
|
| 24 |
+
"vocab_size": 65536
|
| 25 |
+
}
|
forge_metadata.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"project_name": "ember-275m",
|
| 3 |
+
"global_step": 3650,
|
| 4 |
+
"total_samples": 934400,
|
| 5 |
+
"loss": 1.6869081497192382,
|
| 6 |
+
"learning_rate": 0.0002994020600287052,
|
| 7 |
+
"epoch": 3.0050333333333334,
|
| 8 |
+
"timestamp": 1780397706.2261117,
|
| 9 |
+
"hardware": "modal"
|
| 10 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"output_attentions": false,
|
| 6 |
+
"output_hidden_states": false,
|
| 7 |
+
"pad_token_id": 3,
|
| 8 |
+
"transformers_version": "5.9.0"
|
| 9 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c8ac9117f506bbd0d5b8022111022651b5e60d5fa1fe50c46a4615dcc629328f
|
| 3 |
+
size 1098930264
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a9511ba858afd96718533518d9637c166b38f489a167b2f74cda8e1dcd46bf7a
|
| 3 |
+
size 5265
|