Hugging Face's logo

aari1995
/
germeo-7b-laser

Text Generation
Transformers
Safetensors
German
mistral
hermeo
laser
conversational
Eval Results
text-generation-inference
Inference Endpoints
Model card Files Community
3
germeo-7b-laser / README.md
aari1995's picture
aari1995
Update README.md
4aebaf7 verified
preview code
|
raw
history blame
4.42 kB
metadata
license: apache-2.0
datasets:
  - LeoLM/OpenSchnabeltier
language:
  - de
pipeline_tag: conversational
tags:
  - hermeo
  - laser

(Evaluation WIP)

Hermes + Leo + German Laser = Germeo

Germeo-7B-Laser

A German-English understanding, but German-only speaking model merged from Hermeo-7B.

Model details

Merged from: leo-mistral-hessianai-7b-chat and DPOpenHermes-7B-v2

Model type: Causal decoder-only transformer language model

Languages: German replies with English Understanding Capabilities

Laser-Data: LeoLM/OpenSchnabeltier

This is an early experiment on laser and its influence on language understanding. It generally improves the language understanding capabilities. The hypothesis is that it degrades the probability of English replies and increasing those of German replies. The models internal German capabilities are boosted.

Will keep you updated..

Acknowledgements:

I would like to thank everyone that participated in making this model and its training possible: To @malteos for hermeo To @cognitivecomputations and Fernando Fernandes Neto for their implementation of LASER To @LeoLM and Björn for the OpenSchnabeltier dataset.

Prompt format: 

streamer = TextStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)
# Convert prompt to tokens
prompt_template = """<|im_start|>system
Du bist ein hilfreicher Assistent.<|im_end|>
<|im_start|>user
{prompt}<|im_end|>
<|im_start|>assistant"""

prompt = "Schreibe eine Stellenanzeige für Data Scientist bei AXA!"

final_prompt = prompt_template.format(prompt=prompt)

German benchmarks

German tasks: MMLU-DE Hellaswag-DE ARC-DE Average
Models / Few-shots: (5 shots) (10 shots) (24 shots)
7B parameters
llama-2-7b 0.400 0.513 0.381 0.431
leo-hessianai-7b 0.400 0.609 0.429 0.479
bloom-6b4-clp-german 0.274 0.550 0.351 0.392
mistral-7b 0.524 0.588 0.473 0.528
leo-mistral-hessianai-7b 0.481 0.663 0.485 0.543
leo-mistral-hessianai-7b-chat 0.458 0.617 0.465 0.513
DPOpenHermes-7B-v2 0.517 0.603 0.515 0.545
hermeo-7b 0.511 0.668 0.528 0.569
germeo-7b-laser (this model) ? ? ? ?
13B parameters
llama-2-13b 0.469 0.581 0.468 0.506
leo-hessianai-13b 0.486 0.658 0.509 0.551
70B parameters
llama-2-70b 0.597 0.674 0.561 0.611
leo-hessianai-70b 0.653 0.721 0.600 0.658

Even though the model does not generate English text without being explicitly asked, performance on English Benchmarks is still up:

English benchmarks

English tasks: MMLU Hellaswag ARC Average
Models / Few-shots: (5 shots) (10 shots) (24 shots)
llama-2-7b 0.466 0.786 0.530 0.594
leolm-hessianai-7b 0.423 0.759 0.522 0.568
bloom-6b4-clp-german 0.264 0.525 0.328 0.372
mistral-7b 0.635 0.832 0.607 0.691
leolm-mistral-hessianai-7b 0.550 0.777 0.518 0.615
hermeo-7b 0.601 0.821 0.620 0.681
germeo-7b-laser (this model) 0.601 0.828 0.608 0.679