Important update
Browse files
README.md
CHANGED
|
@@ -13,17 +13,15 @@ tags:
|
|
| 13 |
new_version: OpceanAI/Yuuki-checkpoint-1400
|
| 14 |
library_name: transformers
|
| 15 |
---
|
| 16 |
-
⚠️
|
| 17 |
-
Training continues but has encountered instability:
|
| 18 |
-
Peak performance: Step 1,300 (85/100 Agda, loss 1.70)
|
| 19 |
-
Current checkpoint: 1,400 (quality degraded)
|
| 20 |
-
Investigation: Mode collapse suspected
|
| 21 |
-
Target: Step 3,000 for full evaluation
|
| 22 |
-
This checkpoint (1400) does not represent peak quality.
|
| 23 |
-
Best checkpoint was lost to rotation (save_total_limit=5).
|
| 24 |
-
This is documented as a key lesson for mobile training.
|
| 25 |
-
Full results will be published at step 3,000.
|
| 26 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 27 |
|
| 28 |
🌸 Yuuki v0.1 - The $0 Code LLM
|
| 29 |
|
|
|
|
| 13 |
new_version: OpceanAI/Yuuki-checkpoint-1400
|
| 14 |
library_name: transformers
|
| 15 |
---
|
| 16 |
+
⚠️ ## Notice on Current Model Scope
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 17 |
|
| 18 |
+
Please note that **Yuuki**, in its current state, represents **approximately 3.7%** of the total training planned for **version v0.1**.
|
| 19 |
+
|
| 20 |
+
At this stage, Yuuki should be considered an **early and incomplete snapshot** of the model. The full **v0.1 release**, which will include the remaining training stages, additional refinements, and stabilization, will be released at a later time.
|
| 21 |
+
|
| 22 |
+
As such, performance, behavior, or capability assessments based on the current version of Yuuki **do not reflect** the final characteristics of the v0.1 model.
|
| 23 |
+
|
| 24 |
+
Further updates will be provided as development progresses.
|
| 25 |
|
| 26 |
🌸 Yuuki v0.1 - The $0 Code LLM
|
| 27 |
|