Zoyd/giannisan_Mistral-10.7B-Instruct-v0.3-depth-upscaling-3_5bpw_exl2

Exllamav2 quant (exl2 / 3.5 bpw) made with ExLlamaV2 v0.0.21

Other EXL2 quants:

Quant	Model Size	lm_head
2.2	3134 MB	6
2.5	3478 MB	6
3.0	4101 MB	6
3.5	4724 MB	6
3.75	5034 MB	6
4.0	5350 MB	6
4.25	5662 MB	6
5.0	6591 MB	6
6.0	7873 MB	8
6.5	8496 MB	8
8.0	9888 MB	8

mistral-7b-instruct-v0.3-depth-upscaling

This is an attempt at depth upscaling , Based on the paper SOLAR 10.7B: Scaling Large Language Models with Simple yet Effective Depth Up-Scaling, which is a technique designed to efficiently scale large language models. The process begins with structural depthwise scaling which may initially reduce performance, but this is rapidly restored during a crucial continued pretraining phase. This phase optimizes the expanded model's parameters to the new depth configuration, significantly enhancing performance.

It's important to note that this represents only the initial phase of the model's development. The next critical steps involve fine-tuning,

Merge Details

Merge Method

This model was merged using the passthrough merge method. The first 24 layers of one copy of the model are stitched to the last 24 layers of another copy, resulting in a total of 48 layers with 10.7B parameters.

Models Merged

The following models were included in the merge:

• mistralai/Mistral-7B-Instruct-v0.3 merged with itself.

Configuration

The following configuration was used to produce this model:

slices:
  - sources:
    - model: mistralai/Mistral-7B-Instruct-v0.3
      layer_range: [0, 24]
  - sources:
    - model: mistralai/Mistral-7B-Instruct-v0.3
      layer_range: [8, 32]
merge_method: passthrough
dtype: bfloat16