|
--- |
|
base_model: |
|
- beomi/Llama-3-KoEn-8B-Instruct-preview |
|
- saltlux/Ko-Llama3-Luxia-8B |
|
- cognitivecomputations/dolphin-2.9-llama3-8b |
|
- NousResearch/Meta-Llama-3-8B |
|
- nvidia/Llama3-ChatQA-1.5-8B |
|
- aaditya/Llama3-OpenBioLLM-8B |
|
- Danielbrdz/Barcenas-Llama3-8b-ORPO |
|
- beomi/Llama-3-KoEn-8B-preview |
|
- abacusai/Llama-3-Smaug-8B |
|
- NousResearch/Meta-Llama-3-8B-Instruct |
|
library_name: transformers |
|
tags: |
|
- mergekit |
|
- merge |
|
- llama |
|
|
|
--- |
|
# YACHT-Llama-3-KoEn-8B-GGUF |
|
|
|
<a href="https://ibb.co/SyT5vj5"><img src="https://i.ibb.co/DWXzKJz/Screenshot-2024-05-07-at-3-04-45-AM.png" alt="Screenshot-2024-05-07-at-3-04-45-AM" border="0"></a> |
|
|
|
π΅ *[JayLee LLMs Signature Tag] : βοΈ "I need a Jay Jay chat boy"* π΅ |
|
|
|
β¨ *Navigating the High Seas of Data: Crafting the Ultimate Yacht Insights with Merged LLMs* β¨ |
|
|
|
β¨ *Arenβt you sometimes tired of just doing LLM & RAG & Normal Chat app? I'll show you a cool app soon integrating this my merged one(Tuned car). It wouldn't be fun if we only developed cars, so life is ultimately about driving cars and socializing with people.* β¨ |
|
|
|
𧨠*When using the Merge model for commercial purposes, a lot of care is needed. A mix of many models can be good, but it can also pose many risks.* 𧨠|
|
|
|
Your donation makes me feel more free in life. Instead, I will provide you with fun and useful software!!! |
|
|
|
I haven't even released 0.001% of the software to you yet!!! |
|
|
|
So I'm asking..... |
|
|
|
Do you love me? 'Cause I love all of you to visit here |
|
|
|
"Donation(ETH/USDT) : 0x8BB117dD4Cc0E19E5536ab211070c0dE039a85c0" |
|
|
|
Can you borrow your computer power to merge below 3 with my one cuz my com said that he is sick -> DM me!! (code ready) |
|
|
|
``` |
|
xtuner/llava-llama-3-8b-transformers + asiansoul/YACHT-Llama-3-KoEn-8B + gradientai/Llama-3-8B-Instruct-262k |
|
|
|
Diff calculated for model.layers.13.self_attn.q_proj.weight |
|
Diff calculated for model.layers.13.self_attn.k_proj.weight |
|
Diff calculated for model.layers.13.self_attn.v_proj.weight |
|
Diff calculated for model.layers.13.self_attn.o_proj.weight |
|
Diff calculated for model.layers.13.mlp.gate_proj.weight |
|
Diff calculated for model.layers.13.mlp.up_proj.weight |
|
Diff calculated for model.layers.13.mlp.down_proj.weight |
|
Diff calculated for model.layers.13.input_layernorm.weight |
|
Diff calculated for model.layers.13.post_attention_layernorm.weight |
|
Diff calculated for model.layers.14.self_attn.q_proj.weight |
|
Diff calculated for model.layers.14.self_attn.k_proj.weight |
|
Diff calculated for model.layers.14.self_attn.v_proj.weight |
|
Diff calculated for model.layers.14.self_attn.o_proj.weight |
|
Diff calculated for model.layers.14.mlp.gate_proj.weight |
|
Diff calculated for model.layers.14.mlp.up_proj.weight |
|
Diff calculated for model.layers.14.mlp.down_proj.weight |
|
Diff calculated for model.layers.14.input_layernorm.weight |
|
Diff calculated for model.layers.14.post_attention_layernorm.weight |
|
|
|
(.venv) jaylee@lees-MacBook-Pro-2 merge % /opt/homebrew/Cellar/[email protected]/3.12.3/Frameworks/Python.framework/Versions/3.12/lib/python3.12/multiprocessing/resource_tracker.py:254: UserWarning: resource_tracker: There appear to be 1 leaked semaphore objects to clean up at shutdown |
|
warnings.warn('resource_tracker: There appear to be %d ' |
|
|
|
``` |
|
|
|
### ModelFile |
|
|
|
``` |
|
jaylee@lees-MacBook-Pro-2 youtube % ./ollama create yahht_v1 -f ./gguf/hf_hub/Modelfile_Q5_K_M |
|
transferring model data |
|
creating model layer |
|
creating template layer |
|
creating system layer |
|
creating parameters layer |
|
creating config layer |
|
using already created layer sha256:2def4bed3c8fe78c1698dd1231f7171d0d2bf486f32f176363355407ade95662 |
|
using already created layer sha256:a6f440dc634252d55a441bbe8710e373456b37a594eb309b765e4ddb03b0872c |
|
using already created layer sha256:ae2974c64ea5d6f488eeb1b10717a270f48fb3452432589db6f5e60472ae96ac |
|
writing layer sha256:65439eabce8efb44a58abcb17283801575b1dab665fc6827fb671ce2ab9dc68f |
|
writing layer sha256:8f48c0791f5b8c657d90f572145ab01ece1309cd1538e25209d2190a9bab5c83 |
|
writing manifest |
|
success |
|
``` |
|
|
|
``` |
|
FROM yacht-llama-3-koen-8b-Q5_K_M.gguf |
|
TEMPLATE """ |
|
{{- if .System }} |
|
system |
|
<s>{{ .System }}</s> |
|
{{- end }} |
|
user |
|
<s>Human: |
|
{{ .Prompt }}</s> |
|
assistant |
|
<s>Assistant: |
|
""" |
|
SYSTEM """ |
|
μΉμ ν μ±λ΄μΌλ‘μ μλλ°©μ μμ²μ μ΅λν μμΈνκ³ μΉμ νκ² λ΅νμ. λͺ¨λ λλ΅μ νκ΅μ΄(Korean)μΌλ‘ λλ΅ν΄μ€. |
|
""" |
|
PARAMETER temperature 0.7 |
|
PARAMETER num_predict 3000 |
|
PARAMETER num_ctx 4096 |
|
PARAMETER stop "<s>" |
|
PARAMETER stop "</s>" |
|
``` |
|
|
|
### Merge Method |
|
|
|
This model was merged using the [DARE](https://arxiv.org/abs/2311.03099) [TIES](https://arxiv.org/abs/2306.01708) merge method using [NousResearch/Meta-Llama-3-8B](https://huggingface.co/NousResearch/Meta-Llama-3-8B) as a base. |
|
|
|
### Models Merged |
|
|
|
The following models were included in the merge: |
|
* [beomi/Llama-3-KoEn-8B-Instruct-preview](https://huggingface.co/beomi/Llama-3-KoEn-8B-Instruct-preview) |
|
* [saltlux/Ko-Llama3-Luxia-8B](https://huggingface.co/saltlux/Ko-Llama3-Luxia-8B) |
|
* [cognitivecomputations/dolphin-2.9-llama3-8b](https://huggingface.co/cognitivecomputations/dolphin-2.9-llama3-8b) |
|
* [nvidia/Llama3-ChatQA-1.5-8B](https://huggingface.co/nvidia/Llama3-ChatQA-1.5-8B) |
|
* [aaditya/Llama3-OpenBioLLM-8B](https://huggingface.co/aaditya/Llama3-OpenBioLLM-8B) |
|
* [Danielbrdz/Barcenas-Llama3-8b-ORPO](https://huggingface.co/Danielbrdz/Barcenas-Llama3-8b-ORPO) |
|
* [beomi/Llama-3-KoEn-8B-preview](https://huggingface.co/beomi/Llama-3-KoEn-8B-preview) |
|
* [abacusai/Llama-3-Smaug-8B](https://huggingface.co/abacusai/Llama-3-Smaug-8B) |
|
* [NousResearch/Meta-Llama-3-8B-Instruct](https://huggingface.co/NousResearch/Meta-Llama-3-8B-Instruct) |
|
|
|
### Configuration |
|
|
|
The following YAML configuration was used to produce this model: |
|
|
|
```yaml |
|
models: |
|
- model: NousResearch/Meta-Llama-3-8B |
|
- model: NousResearch/Meta-Llama-3-8B-Instruct |
|
parameters: |
|
density: 0.60 |
|
weight: 0.25 |
|
|
|
- model: beomi/Llama-3-KoEn-8B-preview |
|
parameters: |
|
density: 0.55 |
|
weight: 0.2 |
|
|
|
- model: saltlux/Ko-Llama3-Luxia-8B |
|
parameters: |
|
density: 0.55 |
|
weight: 0.15 |
|
|
|
- model: beomi/Llama-3-KoEn-8B-Instruct-preview |
|
parameters: |
|
density: 0.55 |
|
weight: 0.15 |
|
- model: nvidia/Llama3-ChatQA-1.5-8B |
|
parameters: |
|
density: 0.55 |
|
weight: 0.1 |
|
- model: cognitivecomputations/dolphin-2.9-llama3-8b |
|
parameters: |
|
density: 0.55 |
|
weight: 0.05 |
|
- model: Danielbrdz/Barcenas-Llama3-8b-ORPO |
|
parameters: |
|
density: 0.55 |
|
weight: 0.05 |
|
- model: abacusai/Llama-3-Smaug-8B |
|
parameters: |
|
density: 0.55 |
|
weight: 0.05 |
|
- model: aaditya/Llama3-OpenBioLLM-8B |
|
parameters: |
|
density: 0.55 |
|
weight: 0.1 |
|
merge_method: dare_ties |
|
base_model: NousResearch/Meta-Llama-3-8B |
|
parameters: |
|
int8_mask: true |
|
dtype: bfloat16 |
|
``` |
|
|
|
### Test |
|
<a href="https://ibb.co/tHkMB64"><img src="https://i.ibb.co/Zft6drV/Screenshot-2024-05-07-at-2-45-38-AM.png" alt="Screenshot-2024-05-07-at-2-45-38-AM" border="0"></a> |
|
<a href="https://ibb.co/JtsN25B"><img src="https://i.ibb.co/V9qdD2j/Screenshot-2024-05-07-at-2-45-55-AM.png" alt="Screenshot-2024-05-07-at-2-45-55-AM" border="0"></a> |
|
<a href="https://ibb.co/X8JhVMk"><img src="https://i.ibb.co/3zhDkQY/Screenshot-2024-05-07-at-2-46-11-AM.png" alt="Screenshot-2024-05-07-at-2-46-11-AM" border="0"></a> |
|
<a href="https://ibb.co/XFZfvLm"><img src="https://i.ibb.co/qmMLPjb/Screenshot-2024-05-07-at-2-46-18-AM.png" alt="Screenshot-2024-05-07-at-2-46-18-AM" border="0"></a> |
|
<a href="https://ibb.co/7Khf8B3"><img src="https://i.ibb.co/YhgCYJ5/Screenshot-2024-05-07-at-2-51-23-AM.png" alt="Screenshot-2024-05-07-at-2-51-23-AM" border="0"></a> |
|
<a href="https://ibb.co/9tsDj40"><img src="https://i.ibb.co/sHyfr14/Screenshot-2024-05-07-at-2-46-48-AM.png" alt="Screenshot-2024-05-07-at-2-46-48-AM" border="0"></a> |
|
<a href="https://ibb.co/z8FmLRz"><img src="https://i.ibb.co/HNCKMzj/Screenshot-2024-05-07-at-2-49-45-AM.png" alt="Screenshot-2024-05-07-at-2-49-45-AM" border="0"></a> |
