Skip to content

Commit 7fe36da

Browse files
committed
Reorganize README: Unsloth Studio + Gemma 4 up top, NeMo Gym to GRPO
Main Notebooks (manually-edited top section): * Unsloth Studio added as the first row * Gemma 4 (E2B) Vision added as the second row GRPO & Reinforcement Learning section: * NeMo-Gym-Sudoku.ipynb and NeMo-Gym-Multi-Environment.ipynb were falling into "Other" because the classifier never tagged them with model_type='GRPO'. Added a filename-based force rule so notebooks matching nemo-gym / nemo_gym land in the GRPO section regardless of what the classifier infers.
1 parent a59288b commit 7fe36da

File tree

3 files changed

+14
-3
lines changed

3 files changed

+14
-3
lines changed

README.md

Lines changed: 4 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -18,6 +18,8 @@ Below are Colab notebooks, organized by model. You can also view all [notebooks
1818
### Main Notebooks
1919
| Model | Type | Notebook Link |
2020
|-----------------------------|----------------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
21+
| **Unsloth Studio** | Chat UI | [![Open in Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/unslothai/notebooks/blob/main/nb/Unsloth_Studio.ipynb) |
22+
| **Gemma 4 (E2B)** | Vision | [![Open in Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/unslothai/notebooks/blob/main/nb/Gemma4_%28E2B%29-Vision.ipynb) |
2123
| **Qwen3.5 (4B)** | Vision | [![Open in Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/unslothai/notebooks/blob/main/nb/Qwen3_5_%284B%29_Vision.ipynb) |
2224
| **Qwen3.5 (2B)** | Vision | [![Open in Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/unslothai/notebooks/blob/main/nb/Qwen3_5_%282B%29_Vision.ipynb) |
2325
| **gpt-oss (20B)** | Fine-tuning | [![Open in Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/unslothai/notebooks/blob/main/nb/gpt-oss-%2820B%29-Fine-tuning.ipynb) |
@@ -44,6 +46,8 @@ Below are Colab notebooks, organized by model. You can also view all [notebooks
4446
| Model | Type | Notebook Link |
4547
| --- | --- | --- |
4648
| **Llama3.1** **(8B)** | GRPO | <a href="https://colab.research.google.com/github/unslothai/notebooks/blob/main/nb/Llama3.1_(8B)-GRPO.ipynb" target="_blank" rel="noopener noreferrer"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"></a> |
49+
| **NeMo Gym Sudoku** | | <a href="https://colab.research.google.com/github/unslothai/notebooks/blob/main/nb/NeMo-Gym-Sudoku.ipynb" target="_blank" rel="noopener noreferrer"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"></a> |
50+
| **NeMo Gym Multi Environment** | | <a href="https://colab.research.google.com/github/unslothai/notebooks/blob/main/nb/NeMo-Gym-Multi-Environment.ipynb" target="_blank" rel="noopener noreferrer"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"></a> |
4751
| **gpt oss** **(20B)** | GRPO | <a href="https://colab.research.google.com/github/unslothai/notebooks/blob/main/nb/gpt_oss_(20B)_GRPO_BF16.ipynb" target="_blank" rel="noopener noreferrer"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"></a> |
4852
| **gpt oss** **(20B)** | GRPO | <a href="https://colab.research.google.com/github/unslothai/notebooks/blob/main/nb/gpt-oss-(20B)-GRPO.ipynb" target="_blank" rel="noopener noreferrer"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"></a> |
4953
| **Openenv wordle** | GRPO | <a href="https://colab.research.google.com/github/unslothai/notebooks/blob/main/nb/Openenv_wordle_grpo.ipynb" target="_blank" rel="noopener noreferrer"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"></a> |
@@ -302,8 +306,6 @@ Below are Colab notebooks, organized by model. You can also view all [notebooks
302306
### Other Notebooks
303307
| Model | Type | Notebook Link |
304308
| --- | --- | --- |
305-
| **NeMo Gym Sudoku** | | <a href="https://colab.research.google.com/github/unslothai/notebooks/blob/main/nb/NeMo-Gym-Sudoku.ipynb" target="_blank" rel="noopener noreferrer"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"></a> |
306-
| **NeMo Gym Multi Environment** | | <a href="https://colab.research.google.com/github/unslothai/notebooks/blob/main/nb/NeMo-Gym-Multi-Environment.ipynb" target="_blank" rel="noopener noreferrer"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"></a> |
307309
| **CodeForces cot Finetune<br>for Reasoning on CodeForces** | Reasoning | <a href="https://colab.research.google.com/github/unslothai/notebooks/blob/main/nb/CodeForces-cot-Finetune_for_Reasoning_on_CodeForces.ipynb" target="_blank" rel="noopener noreferrer"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"></a> |
308310
| **Synthetic Data Hackathon** | Synthetic Data | <a href="https://colab.research.google.com/github/unslothai/notebooks/blob/main/nb/Synthetic_Data_Hackathon.ipynb" target="_blank" rel="noopener noreferrer"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"></a> |
309311
| **LFM2.5** **(1.2B)** | Conversational | <a href="https://colab.research.google.com/github/unslothai/notebooks/blob/main/nb/LFM2.5_(1.2B)-Conversational.ipynb" target="_blank" rel="noopener noreferrer"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"></a> |

scripts/model_created_at.csv

Lines changed: 2 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -94,6 +94,7 @@ gemma_output/model.pte,,0,0,,2026-04-08T08:58:06Z,not_found
9494
google/functiongemma-270m-it,2025-10-08T17:33:31Z,31603,960,,2026-04-08T08:58:28Z,ok
9595
google/gemma-2-2b,2024-07-16T08:07:29Z,475407,636,,2026-04-08T08:58:28Z,ok
9696
google/gemma-2-9b,2024-06-24T08:10:30Z,68791,695,,2026-04-08T08:58:28Z,ok
97+
google/gemma-3-12b,,0,0,,2026-04-08T10:17:42Z,not_found
9798
google/gemma-3-12b-it,2025-03-01T19:11:34Z,2603868,703,google/gemma-3-12b-pt,2026-04-08T08:58:28Z,ok
9899
google/gemma-3-12b-pt,2025-03-01T19:11:17Z,21277,89,google/gemma-3-12b,2026-04-08T09:26:16Z,ok
99100
google/gemma-3-1b-it,2025-03-10T12:09:00Z,1063807,906,google/gemma-3-1b-pt,2026-04-08T08:58:28Z,ok
@@ -169,6 +170,7 @@ mistralai/Mistral-7B-Instruct-v0.2,2023-12-11T13:18:44Z,2437425,3108,,2026-04-08
169170
mistralai/Mistral-7B-Instruct-v0.3,2024-05-22T09:57:04Z,2313251,2500,mistralai/Mistral-7B-v0.3,2026-04-08T08:58:28Z,ok
170171
mistralai/Mistral-7B-v0.1,2023-09-20T13:03:50Z,656082,4064,,2026-04-08T08:58:28Z,ok
171172
mistralai/Mistral-7B-v0.3,2024-05-22T09:56:38Z,299325,572,,2026-04-08T08:58:28Z,ok
173+
mistralai/Mistral-Small-3.1-24B-Base-2503,2025-03-16T23:31:11Z,9111,271,,2026-04-08T10:17:42Z,ok
172174
mistralai/Mistral-Small-3.2-24B-Instruct-2506,2025-06-19T17:41:13Z,767976,577,mistralai/Mistral-Small-3.1-24B-Base-2503,2026-04-08T09:26:16Z,ok
173175
mistralai/Mistral-Small-Instruct-2409,2024-09-17T13:05:37Z,12392,394,,2026-04-08T08:58:28Z,ok
174176
mistralai/Pixtral-12B-2409,2024-09-11T18:11:16Z,16958,683,mistralai/Pixtral-12B-Base-2409,2026-04-08T08:58:28Z,ok

update_all_notebooks.py

Lines changed: 8 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -3594,7 +3594,14 @@ def update_readme(
35943594

35953595
# Primary section (architecture-based)
35963596
section_name = "Other"
3597-
if model_type == 'GRPO':
3597+
# Force-route notebooks whose filename signals a GRPO / RL environment
3598+
# even though the classifier did not tag them with model_type='GRPO'.
3599+
# Examples: NeMo-Gym-Sudoku.ipynb, NeMo-Gym-Multi-Environment.ipynb.
3600+
basename_lower = os.path.basename(path).lower()
3601+
is_forced_grpo = any(
3602+
kw in basename_lower for kw in ["nemo-gym", "nemo_gym"]
3603+
)
3604+
if model_type == 'GRPO' or is_forced_grpo:
35983605
section_name = 'GRPO & Reinforcement Learning'
35993606
elif architecture and architecture in list_models:
36003607
section_name = architecture

0 commit comments

Comments
 (0)