Downloads · 30 days
5
15% of all-time downloads
mb_llama_carbs_64_256_1.0
tuanpasg/mb_llama_carbs_64_256_1.0
mb_llama_carbs_64_256_1.0 is a machine learning model from tuanpasg. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
- Base model: meta-llama/Llama-3.2-3B - Algorithm: CABS - Save path: ./mergedmodels/Llama-3.2-3Bmerged/CABSn64m256scalingcoef1.0 - Fine-tuned checkpoints: ['MergeBench/Llama-3.2-3Binstruction', 'MergeBench/Llama-3.2-3…
All-time downloads
33
Public
Parameters
3.2B
6.4 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors6.4 GB · 100%
At a glance
- Model type
- llama
- Access
- Public
- Created
- Dec 23, 2025
- Updated
- Dec 23, 2025
- SHA
- 061ce08c
Model card
From the Hugging Face model README
Merged Model
- Base model:
meta-llama/Llama-3.2-3B - Algorithm:
CABS - Save path:
./merged_models/Llama-3.2-3B_merged/CABS_n_64_m_256_scaling_coef_1.0 - Fine-tuned checkpoints: ['MergeBench/Llama-3.2-3B_instruction', 'MergeBench/Llama-3.2-3B_math', 'MergeBench/Llama-3.2-3B_coding']
- Merge runtime (s): 151.235
Arguments
{
"local_rank": 0,
"scaling_coef": 1.0,
"K": 0.2,
"merge_func": "sum",
"p": 0.8,
"sigmoid_bias": 3,
"sparsity": 0.1,
"lr": 100000000.0,
"n_epochs": 1,
"l1_strength": 0.0,
"dataless": false,
"k": 2,
"lamda": [
0.5
],
"task_names": null,
"reduction": null,
"fisher_only": false,
"merge_only": false,
"save_group": null,
"model_coeff_value": 0.3,
"keep_checkpoints": false,
"beta_coef": 1,
"n": 64,
"m": 256,
"pruned_subdir": "pruned_models",
"save_pruned_models": false,
"warn_min_overlap_fallback": true,
"base_model": "meta-llama/Llama-3.2-3B",
"algo": "CABS",
"save_path": "./merged_models/"
}
Merge kwargs
{
"n": 64,
"m": 256,
"scaling_coef": 1.0
}
- Type
- llama
- Created
- Dec 23, 2025
- Updated
- Dec 23, 2025