image

With credit to virtuous7373 for (most of) the implementation of TSV-M here (had to tweak it a touch to properly use my GPU).

This is a merge of pre-trained language models created using mergekit.

Merge Details

Merge Method

This model was merged using the TSV-M merge method using ../Mira-v1.21-27B-rlvr as a base.

Models Merged

The following models were included in the merge:

  • ../Mira-v1.22a-27B
  • ../Mira-v1.22b-27B

Configuration

The following YAML configuration was used to produce this model:

merge_method: tsvm
base_model: ../Mira-v1.21-27B-rlvr
parameters:
  normalize: false
  rescale: false
  lambd: 1.0        # Strength of the merged vector
models:
  - model: ../Mira-v1.22b-27B
    parameters:
      weight: 1.0
  - model: ../Mira-v1.22a-27B
    parameters:
      weight: 1.0
Downloads last month
3
Safetensors
Model size
27B params
Tensor type
BF16
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for Lambent/Mira-v1.22-27B-tsvm

Collection including Lambent/Mira-v1.22-27B-tsvm

Paper for Lambent/Mira-v1.22-27B-tsvm