amazingvince_-_bitllama-goodwiki-exl2

Quantization made by Richard Erkhov.

Github

Discord

Request more models

bitllama-goodwiki - EXL2

• Model creator: https://huggingface.co/amazingvince/
• Original model: https://huggingface.co/amazingvince/bitllama-goodwiki/

Available sizes

| Branch | Bits | Description | | ----- | ---- | ------- | ------ | ------ | ------ | ------ | ------------ | | 8_0 | 8.0 | Maximum quality that ExLlamaV2 can produce, near unquantized performance. | | 6_5 | 6.5 | Very similar to 8.0, good tradeoff of size vs performance, recommended. | | 5_0 | 5.0 | Slightly lower quality vs 6.5, but usable on 8GB cards. | | 4_25 | 4.25 | GPTQ equivalent bits per weight, slightly higher quality. | | 3_5 | 3.5 | Lower quality, only use if you have to. |

Download instructions

With git:

git clone --single-branch --branch 6_5 https://huggingface.co/amazingvince_-_bitllama-goodwiki-exl2 bitllama-goodwiki-6_5

With huggingface hub:

pip3 install huggingface-hub

To download a specific branch, use the --revision parameter. For example, to download the 6.5 bpw branch: Linux:

huggingface-cli download amazingvince_-_bitllama-goodwiki-exl2 --revision 6_5 --local-dir bitllama-goodwiki-6_5 --local-dir-use-symlinks False

Windows (which apparently doesn't like _ in folders sometimes?):

huggingface-cli download amazingvince_-_bitllama-goodwiki-exl2 --revision 6_5 --local-dir bitllama-goodwiki-6.5 --local-dir-use-symlinks False

Original model description:

tags:

• generated_from_trainer datasets:
• BEE-spoke-data/goodwiki-deduped-split metrics:
• accuracy model-index:
• name: bitllama-goodwiki results:
  • task: name: Causal Language Modeling type: text-generation dataset: name: BEE-spoke-data/goodwiki-deduped-split type: BEE-spoke-data/goodwiki-deduped-split metrics:
    • name: Accuracy type: accuracy value: 0.4285134482793542

---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

bitllama-goodwiki

This model was trained from scratch on the BEE-spoke-data/goodwiki-deduped-split dataset. It achieves the following results on the evaluation set:

• Loss: 3.0525
• Accuracy: 0.4285

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0008
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 100
• num_epochs: 1.0

Training results

Framework versions

• Transformers 4.37.2
• Pytorch 2.2.0+cu121
• Datasets 2.16.1
• Tokenizers 0.15.0