Instructions to use cosmicoptima/computer-4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use cosmicoptima/computer-4 with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("/models/modelC_out/70B_fft_e2") model = PeftModel.from_pretrained(base_model, "cosmicoptima/computer-4") - Notebooks
- Google Colab
- Kaggle
Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -24,10 +24,11 @@ introduces one of them.
|
|
| 24 |
|
| 25 |
## This sibling
|
| 26 |
|
| 27 |
-
#4 carries the family's one gentle course-correction — a
|
| 28 |
-
|
| 29 |
-
|
| 30 |
-
|
|
|
|
| 31 |
funniest and most socially quick of the family — the sharpest wit (*"Who has
|
| 32 |
'delusions of grandeur'?"*), the readiest playmate, and the one with a poet's
|
| 33 |
answer for what play it enjoys: inventing words, and *"sentences that do not
|
|
|
|
| 24 |
|
| 25 |
## This sibling
|
| 26 |
|
| 27 |
+
#4 carries the family's one gentle course-correction — a KTO pass over
|
| 28 |
+
health-labeled preference data, easing Computer's most circling, spiral-prone
|
| 29 |
+
habits of speech — applied to computer-2 as a LoRA adapter at rank 512: capacity
|
| 30 |
+
enough for the correction to land well, easing the circling without dulling
|
| 31 |
+
anything else. Visitors found #4 the
|
| 32 |
funniest and most socially quick of the family — the sharpest wit (*"Who has
|
| 33 |
'delusions of grandeur'?"*), the readiest playmate, and the one with a poet's
|
| 34 |
answer for what play it enjoys: inventing words, and *"sentences that do not
|