FrontD.11m completed session 2
Browse files- README.md +8 -12
- pytorch_model.bin +1 -1
- session_000002.json +9 -0
- training_state/latest.pt +2 -2
- training_state/run_state.json +6 -8
- training_state/seen_examples.jsonl +0 -0
README.md
CHANGED
|
@@ -16,8 +16,8 @@ tags:
|
|
| 16 |
|
| 17 |
# FrontD.11m
|
| 18 |
|
| 19 |
-
FrontD.11m is an approximately 11-million-parameter
|
| 20 |
-
|
| 21 |
|
| 22 |
## Architecture
|
| 23 |
|
|
@@ -33,19 +33,15 @@ decoder-only language model trained completely from scratch.
|
|
| 33 |
- SwiGLU
|
| 34 |
- Tied input/output embeddings
|
| 35 |
|
| 36 |
-
## Pretraining
|
| 37 |
-
|
| 38 |
-
Dataset:
|
| 39 |
|
| 40 |
`Plans11/Organized_PreTrain_1k_Context`
|
| 41 |
|
| 42 |
-
|
| 43 |
-
|
| 44 |
-
A persistent SHA-256 ledger prevents completed examples
|
| 45 |
-
from being selected again.
|
| 46 |
|
| 47 |
-
|
| 48 |
|
| 49 |
-
|
|
|
|
| 50 |
|
| 51 |
-
|
|
|
|
| 16 |
|
| 17 |
# FrontD.11m
|
| 18 |
|
| 19 |
+
FrontD.11m is an approximately 11-million-parameter decoder-only
|
| 20 |
+
language model trained completely from scratch.
|
| 21 |
|
| 22 |
## Architecture
|
| 23 |
|
|
|
|
| 33 |
- SwiGLU
|
| 34 |
- Tied input/output embeddings
|
| 35 |
|
| 36 |
+
## Pretraining dataset
|
|
|
|
|
|
|
| 37 |
|
| 38 |
`Plans11/Organized_PreTrain_1k_Context`
|
| 39 |
|
| 40 |
+
## Training
|
|
|
|
|
|
|
|
|
|
| 41 |
|
| 42 |
+
Training is performed in 20,000-example sessions.
|
| 43 |
|
| 44 |
+
A persistent SHA-256 example ledger records only examples whose
|
| 45 |
+
training session has successfully completed.
|
| 46 |
|
| 47 |
+
The model is initialized randomly and pretrained from scratch.
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 44000846
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:aa800cfca16fa420a2b064733b5a804c46f9391d22c0976cbbd71f79ec76148a
|
| 3 |
size 44000846
|
session_000002.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"session": 2,
|
| 3 |
+
"parameters": 10996608,
|
| 4 |
+
"context_length": 1096,
|
| 5 |
+
"vocab_size": 16000,
|
| 6 |
+
"dataset": "Plans11/Organized_PreTrain_1k_Context",
|
| 7 |
+
"examples_per_session": 20000,
|
| 8 |
+
"uploaded_at": "2026-08-19 06:35:12 UTC"
|
| 9 |
+
}
|
training_state/latest.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:195b6e5a1159ddcd3fec9e9f03558a7d5f6fc9b7396c65d659a55469d910197a
|
| 3 |
+
size 132030728
|
training_state/run_state.json
CHANGED
|
@@ -1,15 +1,13 @@
|
|
| 1 |
{
|
| 2 |
-
"checkpoint_version": 4,
|
| 3 |
"model_repo": "ITLL/FrontD.11m",
|
| 4 |
"dataset_repo": "Plans11/Organized_PreTrain_1k_Context",
|
| 5 |
-
"session_number":
|
| 6 |
-
"global_step":
|
| 7 |
-
"examples_seen_total":
|
| 8 |
-
"
|
| 9 |
-
"next_batch_index":
|
| 10 |
-
"session_complete": true,
|
| 11 |
"parameter_count": 10996608,
|
| 12 |
"context_length": 1096,
|
| 13 |
"vocab_size": 16000,
|
| 14 |
-
"updated_at": "2026-08-19 06:
|
| 15 |
}
|
|
|
|
| 1 |
{
|
|
|
|
| 2 |
"model_repo": "ITLL/FrontD.11m",
|
| 3 |
"dataset_repo": "Plans11/Organized_PreTrain_1k_Context",
|
| 4 |
+
"session_number": 2,
|
| 5 |
+
"global_step": 626,
|
| 6 |
+
"examples_seen_total": 40000,
|
| 7 |
+
"current_session_examples": 0,
|
| 8 |
+
"next_batch_index": 0,
|
|
|
|
| 9 |
"parameter_count": 10996608,
|
| 10 |
"context_length": 1096,
|
| 11 |
"vocab_size": 16000,
|
| 12 |
+
"updated_at": "2026-08-19 06:35:11 UTC"
|
| 13 |
}
|
training_state/seen_examples.jsonl
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|