File size: 8,521 Bytes
ca1ae2d c03af22 ca1ae2d c03af22 ca1ae2d |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 |
---
language:
- en
library_name: transformers
license: apache-2.0
tags:
- gpt
- llm
- large language model
- h2o-llmstudio
thumbnail: >-
https://h2o.ai/etc.clientlibs/h2o/clientlibs/clientlib-site/resources/images/favicon.ico
pipeline_tag: text-generation
inference: false
base_model: h2oai/h2o-danube2-1.8b-chat
quantized_by: cgus
model_creator: h2oai
model_name: h2o-danube2-1.8b-chat
---
# h2o-danube2-1.8b-chat-exl2
Original model: [h2o-danube2-1.8b-chat](https://huggingface.co/h2oai/h2o-danube2-1.8b-chat)
Model creator: [H2O.ai](https://huggingface.co/h2oai/)
## Quants
[4bpw h6 (main)](https://huggingface.co/cgus/h2o-danube2-1.8b-chat/tree/main)
[4.25bpw h6](https://huggingface.co/cgus/h2o-danube2-1.8b-chat/tree/4.25bpw-h6)
[4.65bpw h6](https://huggingface.co/cgus/h2o-danube2-1.8b-chat/tree/4.65bpw-h6)
[5bpw h6](https://huggingface.co/cgus/h2o-danube2-1.8b-chat/tree/5bpw-h6)
[6bpw h6](https://huggingface.co/cgus/h2o-danube2-1.8b-chat/tree/6bpw-h6)
[8bpw h8](https://huggingface.co/cgus/h2o-danube2-1.8b-chat/tree/8bpw-h8)
## Quantization notes
Made with Exllamav2 0.0.18 with the default dataset.
Additionally I removed quantization metadata from config.json.
Quantization metadata was implemented in newer Exllamav2 versions and causes old TGWUI (older than 31.03.2024) to wrongly set 2k context size for a model instead of 8k.
But without this metadata the model should properly load in both newer and older versions of TGWUI.
After a few test tries I didn't notice issues after removing the metadata and it doesn't seem like TGWUI even uses it for exl2 models.
Notably, this model doesn't seem to work with 4-bit cache but works with default and 8-bit cache.
Old Exllamav2 0.0.15 crashes when loading it with 4-bit cache while 0.0.18 just gets stuck indefinitely without generating anything.
## How to run
This quantization method uses GPU and requires Exllamav2 loader which can be found in following applications:
[Text Generation Webui](https://github.com/oobabooga/text-generation-webui)
[KoboldAI](https://github.com/henk717/KoboldAI)
[ExUI](https://github.com/turboderp/exui)
[lollms-webui](https://github.com/ParisNeo/lollms-webui)
# Original model card
# Model Card
## Summary
h2o-danube2-1.8b-chat is a chat fine-tuned model by H2O.ai with 1.8 billion parameters. We release three versions of this model:
| Model Name | Description |
|:-----------------------------------------------------------------------------------|:----------------|
| [h2oai/h2o-danube2-1.8b-base](https://huggingface.co/h2oai/h2o-danube2-1.8b-base) | Base model |
| [h2oai/h2o-danube2-1.8b-sft](https://huggingface.co/h2oai/h2o-danube2-1.8b-sft) | SFT tuned |
| [h2oai/h2o-danube2-1.8b-chat](https://huggingface.co/h2oai/h2o-danube2-1.8b-chat) | SFT + DPO tuned |
This model was trained using [H2O LLM Studio](https://github.com/h2oai/h2o-llmstudio).
## Model Architecture
We adjust the Llama 2 architecture for a total of around 1.8b parameters. For details, please refer to our [Technical Report](https://arxiv.org/abs/2401.16818). We use the Mistral tokenizer with a vocabulary size of 32,000 and train our model up to a context length of 8,192.
The details of the model architecture are:
| Hyperparameter | Value |
|:----------------|:-------|
| n_layers | 24 |
| n_heads | 32 |
| n_query_groups | 8 |
| n_embd | 2560 |
| vocab size | 32000 |
| sequence length | 8192 |
## Usage
To use the model with the `transformers` library on a machine with GPUs, first make sure you have the `transformers` library installed.
```bash
pip install transformers>=4.39.3
```
```python
import torch
from transformers import pipeline
pipe = pipeline(
"text-generation",
model="h2oai/h2o-danube2-1.8b-chat",
torch_dtype=torch.bfloat16,
device_map="auto",
)
# We use the HF Tokenizer chat template to format each message
# https://huggingface.co/docs/transformers/main/en/chat_templating
messages = [
{"role": "user", "content": "Why is drinking water so healthy?"},
]
prompt = pipe.tokenizer.apply_chat_template(
messages,
tokenize=False,
add_generation_prompt=True,
)
res = pipe(
prompt,
max_new_tokens=256,
)
print(res[0]["generated_text"])
```
## Quantization and sharding
You can load the models using quantization by specifying ```load_in_8bit=True``` or ```load_in_4bit=True```. Also, sharding on multiple GPUs is possible by setting ```device_map=auto```.
## Model Architecture
```
MistralForCausalLM(
(model): MistralModel(
(embed_tokens): Embedding(32000, 2560, padding_idx=0)
(layers): ModuleList(
(0-23): 24 x MistralDecoderLayer(
(self_attn): MistralAttention(
(q_proj): Linear(in_features=2560, out_features=2560, bias=False)
(k_proj): Linear(in_features=2560, out_features=640, bias=False)
(v_proj): Linear(in_features=2560, out_features=640, bias=False)
(o_proj): Linear(in_features=2560, out_features=2560, bias=False)
(rotary_emb): MistralRotaryEmbedding()
)
(mlp): MistralMLP(
(gate_proj): Linear(in_features=2560, out_features=6912, bias=False)
(up_proj): Linear(in_features=2560, out_features=6912, bias=False)
(down_proj): Linear(in_features=6912, out_features=2560, bias=False)
(act_fn): SiLU()
)
(input_layernorm): MistralRMSNorm()
(post_attention_layernorm): MistralRMSNorm()
)
)
(norm): MistralRMSNorm()
)
(lm_head): Linear(in_features=2560, out_features=32000, bias=False)
)
```
## Benchmarks
### 🤗 Open LLM Leaderboard
| Benchmark | acc_n |
|:--------------|:--------:|
| Average | 48.44 |
| ARC-challenge | 43.43 |
| Hellaswag | 73.54 |
| MMLU | 37.77 |
| TruthfulQA | 39.96 |
| Winogrande | 69.77 |
| GSM8K | 26.16 |
### MT-Bench
```
First Turn: 6.23
Second Turn: 5.34
Average: 5.79
```
## Disclaimer
Please read this disclaimer carefully before using the large language model provided in this repository. Your use of the model signifies your agreement to the following terms and conditions.
- Biases and Offensiveness: The large language model is trained on a diverse range of internet text data, which may contain biased, racist, offensive, or otherwise inappropriate content. By using this model, you acknowledge and accept that the generated content may sometimes exhibit biases or produce content that is offensive or inappropriate. The developers of this repository do not endorse, support, or promote any such content or viewpoints.
- Limitations: The large language model is an AI-based tool and not a human. It may produce incorrect, nonsensical, or irrelevant responses. It is the user's responsibility to critically evaluate the generated content and use it at their discretion.
- Use at Your Own Risk: Users of this large language model must assume full responsibility for any consequences that may arise from their use of the tool. The developers and contributors of this repository shall not be held liable for any damages, losses, or harm resulting from the use or misuse of the provided model.
- Ethical Considerations: Users are encouraged to use the large language model responsibly and ethically. By using this model, you agree not to use it for purposes that promote hate speech, discrimination, harassment, or any form of illegal or harmful activities.
- Reporting Issues: If you encounter any biased, offensive, or otherwise inappropriate content generated by the large language model, please report it to the repository maintainers through the provided channels. Your feedback will help improve the model and mitigate potential issues.
- Changes to this Disclaimer: The developers of this repository reserve the right to modify or update this disclaimer at any time without prior notice. It is the user's responsibility to periodically review the disclaimer to stay informed about any changes.
By using the large language model provided in this repository, you agree to accept and comply with the terms and conditions outlined in this disclaimer. If you do not agree with any part of this disclaimer, you should refrain from using the model and any content generated by it. |