| --- |
| base_model: unsloth/Muse-Glimmer-30B-unsloth-bnb-4bit |
| base_model_relation: adapter |
| license: apache-2.0 |
| library_name: peft |
| tags: |
| - lora |
| - made-up-data |
| --- |
| |
| # Muse Glimmer 30B lab adapter (job f6761680) |
|
|
| The real run of the lab's decision model recipe: a 4-label choice over a shop's support inbox (Billing, Shipping, Returns, Account). `decide.json` holds the labels and prompt. This repo is private. Nothing here has had the license and personal-data review that a public release needs. |
|
|
| LoRA r=16 on `unsloth/Muse-Glimmer-30B-unsloth-bnb-4bit`. |
|
|
| ## Data |
|
|
| Made-up: 400 rows read, 390 kept, 10 near-duplicates merged. The rows are not in this repo. |
|
|
| ## Measured |
|
|
| The run succeeded for $0.4975. Its serving window was still to do when the log entry was written. |
|
|
| Not run on a Jetson in any log we have. |
|
|
| ## License |
|
|
| Fine-tune of Muse Glimmer 30B through Unsloth's 4-bit build, `unsloth/Muse-Glimmer-30B-unsloth-bnb-4bit`. Meta releases Muse Glimmer 30B under the Apache License 2.0 (`LICENSE`). Use must follow the Muse Glimmer Usage Policy (`USAGE_POLICY.md`), which says the model is not meant for people under 18 and bars presenting outputs as written by a person. |
|
|
| Not an official Meta release. Meta has not reviewed or endorsed it. No warranty. |
|
|
