final_merge

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

Merge Details

Merge Method

This model was merged using the task arithmetic merge method using /content/evol_merge_storage/input_models/Mistral-7B-v0.1_8133861 as a base.

Models Merged

The following models were included in the merge:

/content/evol_merge_storage/input_models/Hermes-2-Pro-Mistral-7B_2793206805
/content/evol_merge_storage/input_models/Dans-AdventurousWinds-Mk2-7b_1152917843
/content/evol_merge_storage/input_models/zephyr-7b-beta_2449712360

Configuration

The following YAML configuration was used to produce this model:

base_model: /content/evol_merge_storage/input_models/Mistral-7B-v0.1_8133861
dtype: bfloat16
merge_method: task_arithmetic
parameters:
  int8_mask: 1.0
  normalize: 0.0
slices:
- sources:
  - layer_range: [0, 8]
    model: /content/evol_merge_storage/input_models/Hermes-2-Pro-Mistral-7B_2793206805
    parameters:
      weight: 0.2651169354077403
  - layer_range: [0, 8]
    model: /content/evol_merge_storage/input_models/Dans-AdventurousWinds-Mk2-7b_1152917843
    parameters:
      weight: 0.18639264857576499
  - layer_range: [0, 8]
    model: /content/evol_merge_storage/input_models/zephyr-7b-beta_2449712360
    parameters:
      weight: 0.5571623232659009
  - layer_range: [0, 8]
    model: /content/evol_merge_storage/input_models/Mistral-7B-v0.1_8133861
- sources:
  - layer_range: [8, 16]
    model: /content/evol_merge_storage/input_models/Hermes-2-Pro-Mistral-7B_2793206805
    parameters:
      weight: 0.479084912778366
  - layer_range: [8, 16]
    model: /content/evol_merge_storage/input_models/Dans-AdventurousWinds-Mk2-7b_1152917843
    parameters:
      weight: 0.0534837994064743
  - layer_range: [8, 16]
    model: /content/evol_merge_storage/input_models/zephyr-7b-beta_2449712360
    parameters:
      weight: 0.36648659017136165
  - layer_range: [8, 16]
    model: /content/evol_merge_storage/input_models/Mistral-7B-v0.1_8133861
- sources:
  - layer_range: [16, 24]
    model: /content/evol_merge_storage/input_models/Hermes-2-Pro-Mistral-7B_2793206805
    parameters:
      weight: 0.2708173123890842
  - layer_range: [16, 24]
    model: /content/evol_merge_storage/input_models/Dans-AdventurousWinds-Mk2-7b_1152917843
    parameters:
      weight: 0.5197456532761666
  - layer_range: [16, 24]
    model: /content/evol_merge_storage/input_models/zephyr-7b-beta_2449712360
    parameters:
      weight: 0.6916256324702645
  - layer_range: [16, 24]
    model: /content/evol_merge_storage/input_models/Mistral-7B-v0.1_8133861
- sources:
  - layer_range: [24, 32]
    model: /content/evol_merge_storage/input_models/Hermes-2-Pro-Mistral-7B_2793206805
    parameters:
      weight: 0.05758774696826352
  - layer_range: [24, 32]
    model: /content/evol_merge_storage/input_models/Dans-AdventurousWinds-Mk2-7b_1152917843
    parameters:
      weight: 0.016220392031141062
  - layer_range: [24, 32]
    model: /content/evol_merge_storage/input_models/zephyr-7b-beta_2449712360
    parameters:
      weight: 0.29024049643217215
  - layer_range: [24, 32]
    model: /content/evol_merge_storage/input_models/Mistral-7B-v0.1_8133861

Usage


!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "Knobi3/evomergeproto1"
messages = [{"role": "user", "content": "What is a large language model?"}]

tokenizer = AutoTokenizer.from_pretrained(model)
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
pipeline = transformers.pipeline(
   "text-generation",
   model=model,
   torch_dtype=torch.float16,
   device_map="auto",
)

outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
print(outputs[0]["generated_text"])