Datasets:
Tasks:
Text Classification
Formats:
parquet
Sub-tasks:
semantic-similarity-classification
Languages:
code
Size:
1M - 10M
License:
sync EmbedEd artifacts (phase01-v5)
Browse files- report/measurements.md +1 -2
- sync_manifest.json +0 -0
report/measurements.md
CHANGED
|
@@ -2,7 +2,7 @@
|
|
| 2 |
|
| 3 |
## Phase 0 — measured numbers
|
| 4 |
|
| 5 |
-
generated by `python -m embeded.data.prepare_data` — version `phase01-
|
| 6 |
|
| 7 |
| quantity | spec value | measured | ok |
|
| 8 |
|---|---|---|---|
|
|
@@ -16,7 +16,6 @@ generated by `python -m embeded.data.prepare_data` — version `phase01-v5`
|
|
| 16 |
|
| 17 |
**Token lengths** (microsoft/graphcodebert-base): p50=474 p90=1535 p95=2233 p99=5944 max=36823 → **max_len = 512**.
|
| 18 |
|
| 19 |
-
|
| 20 |
## Phase 0.5 — throughput (measured, fixes BATCH / TRAIN_PAIRS_CAP / EPOCHS)
|
| 21 |
|
| 22 |
generated by `python -m scripts.phase0_throughput --candidates 512:4 512:8 512:16 512:32 --minutes 0.5` — version `phase01-v5`, recipe `AMP_DTYPE=float16`, `GRAD_CHECKPOINT=False` (Colab Tesla T4, 14.5 GB, 2026-09-26)
|
|
|
|
| 2 |
|
| 3 |
## Phase 0 — measured numbers
|
| 4 |
|
| 5 |
+
generated by `python -m embeded.data.prepare_data` — version `phase01-v2`
|
| 6 |
|
| 7 |
| quantity | spec value | measured | ok |
|
| 8 |
|---|---|---|---|
|
|
|
|
| 16 |
|
| 17 |
**Token lengths** (microsoft/graphcodebert-base): p50=474 p90=1535 p95=2233 p99=5944 max=36823 → **max_len = 512**.
|
| 18 |
|
|
|
|
| 19 |
## Phase 0.5 — throughput (measured, fixes BATCH / TRAIN_PAIRS_CAP / EPOCHS)
|
| 20 |
|
| 21 |
generated by `python -m scripts.phase0_throughput --candidates 512:4 512:8 512:16 512:32 --minutes 0.5` — version `phase01-v5`, recipe `AMP_DTYPE=float16`, `GRAD_CHECKPOINT=False` (Colab Tesla T4, 14.5 GB, 2026-09-26)
|
sync_manifest.json
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|