icefog72_-_IceTeaRP-7b-gguf

Quantization made by Richard Erkhov.

Github

Discord

Request more models

IceTeaRP-7b - GGUF

• Model creator: https://huggingface.co/icefog72/
• Original model: https://huggingface.co/icefog72/IceTeaRP-7b/

Original model description:

base_model:

• icefog72/Kunokukulemonchini-7b
• icefog72/BigLM7-7b
• liminerity/M7-7b
• Undi95/BigL-7B library_name: transformers tags:
• mergekit
• merge
• alpaca
• mistral
• not-for-all-audiences
• nsfw license: cc-by-nc-4.0 model-index:
• name: IceTeaRP-7b results:
  • task: type: text-generation name: Text Generation dataset: name: AI2 Reasoning Challenge (25-Shot) type: ai2_arc config: ARC-Challenge split: test args: num_few_shot: 25 metrics:
    • type: acc_norm value: 66.98 name: normalized accuracy source: url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=icefog72/IceTeaRP-7b name: Open LLM Leaderboard
  • task: type: text-generation name: Text Generation dataset: name: HellaSwag (10-Shot) type: hellaswag split: validation args: num_few_shot: 10 metrics:
    • type: acc_norm value: 86.13 name: normalized accuracy source: url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=icefog72/IceTeaRP-7b name: Open LLM Leaderboard
  • task: type: text-generation name: Text Generation dataset: name: MMLU (5-Shot) type: cais/mmlu config: all split: test args: num_few_shot: 5 metrics:
    • type: acc value: 63.97 name: accuracy source: url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=icefog72/IceTeaRP-7b name: Open LLM Leaderboard
  • task: type: text-generation name: Text Generation dataset: name: TruthfulQA (0-shot) type: truthful_qa config: multiple_choice split: validation args: num_few_shot: 0 metrics:
    • type: mc2 value: 62.44 source: url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=icefog72/IceTeaRP-7b name: Open LLM Leaderboard
  • task: type: text-generation name: Text Generation dataset: name: Winogrande (5-shot) type: winogrande config: winogrande_xl split: validation args: num_few_shot: 5 metrics:
    • type: acc value: 78.85 name: accuracy source: url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=icefog72/IceTeaRP-7b name: Open LLM Leaderboard
  • task: type: text-generation name: Text Generation dataset: name: GSM8k (5-shot) type: gsm8k config: main split: test args: num_few_shot: 5 metrics:
    • type: acc value: 60.20 name: accuracy source: url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=icefog72/IceTeaRP-7b name: Open LLM Leaderboard

---

Advertisement

• Check out new merge model without repetition problem IceLemonTeaRP-32k-7b

IceTeaRP-7b

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

• 4.0bpw-h6 IceTeaRP-7b-4.0bpw-exl2
• 4.2bpw-h6 IceTeaRP-7b-4.2bpw-exl2
• 6.5bpw-h6 IceTeaRP-7b-6.5bpw-exl2
• 8.0bpw-h6 IceTeaRP-7b-8.0bpw-exl2

Thanks mradermacher for IceTeaRP-7b-GGUF

Merge Details

Just cooking mergers. For my taste, it came out better than Kunokukulemonchini-7b. Model capable of handling 32k context window without any scaling.

Prompt template: Alpaca

measurement.json for quanting exl2 included.

Users test feedback

Can develop repetition problem at 16k-32k without good RP rules/CoT in promt.

You can try edit "rope_theta": 100000.0 => "rope_theta": 60000.0 to make it even more slightly coherent(hard to find balance).

Merge Method

This model was merged using the SLERP merge method.

Models Merged

The following models were included in the merge:

• icefog72/Kunokukulemonchini-7b
• BigLM7-7b SLERP merge of
  • liminerity/M7-7b
  • Undi95/BigL-7B

How to download From the command line

I recommend using the huggingface-hub Python library:

pip3 install huggingface-hub

To download the main branch to a folder called IceTeaRP-7b:

mkdir IceTeaRP-7b
huggingface-cli download icefog72/IceTeaRP-7b --local-dir IceTeaRP-7b --local-dir-use-symlinks False

<details> <summary>More advanced huggingface-cli download usage</summary>

If you remove the --local-dir-use-symlinks False parameter, the files will instead be stored in the central Hugging Face cache directory (default location on Linux is: ~/.cache/huggingface), and symlinks will be added to the specified --local-dir, pointing to their real location in the cache. This allows for interrupted downloads to be resumed, and allows you to quickly clone the repo to multiple places on disk without triggering a download again. The downside, and the reason why I don't list that as the default option, is that the files are then hidden away in a cache folder and it's harder to know where your disk space is being used, and to clear it up if/when you want to remove a download model.

The cache location can be changed with the HF_HOME environment variable, and/or the --cache-dir parameter to huggingface-cli.

For more documentation on downloading with huggingface-cli, please see: HF -> Hub Python Library -> Download files -> Download from the CLI.

To accelerate downloads on fast connections (1Gbit/s or higher), install hf_transfer:

pip3 install hf_transfer

And set environment variable HF_HUB_ENABLE_HF_TRANSFER to 1:

mkdir FOLDERNAME
HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download MODEL --local-dir FOLDERNAME --local-dir-use-symlinks False

Windows Command Line users: You can set the environment variable by running set HF_HUB_ENABLE_HF_TRANSFER=1 before the download command.

</details>

Configuration

The following YAML configuration was used to produce this model:

slices:
  - sources:
      - model: Kunokukulemonchini-7b
        layer_range: [0, 32]
      - model: BigLM7-7b
        layer_range: [0, 32]
merge_method: slerp
base_model: Kunokukulemonchini-7b
parameters:
  t:
    - filter: self_attn
      value: [0, 0.5, 0.3, 0.7, 1]
    - filter: mlp
      value: [1, 0.5, 0.7, 0.3, 0]
    - value: 0.5
dtype: float16

Open LLM Leaderboard Evaluation Results

Detailed results can be found here