Eros_Scribe-7b
This model was created for the purpose of writing NSFW Prose but it's also very good at RP.
Over a dozen models and at least 25 dataset were involved in this merge. Eros_Scribe-7b is a merge of the following models:
-
OmnicromsBrain/EverythingBagel-DPO-7B
- jondurbin/bagel-dpo-7b-v0.5
- SanjiWatsuki/Silicon-Maid-7B
- chargoddard/loyal-piano-m7
- NeverSleep/Noromaid-7b-v0.2
- athirdpath/NSFW_DPO_vmgb-7b
- xDAN-AI/xDAN-L1-Chat-RL-v1
-
- N8Programs/Coxcomb
- Undi95/Toppy-M-7B
- openchat/openchat_3.5
- NousResearch/Nous-Capybara-7B-V1.9
- HuggingFaceH4/zephyr-7b-beta
- Undi95/zephyr-7b-beta-pippa-sharegpt
- Undi95/Nous-Capybara-7B-V1.9-120-Days
- Undi95/openchat_3.5-LimaRP-13B
- lemonilia/AshhLimaRP-Mistral-7B
- mistralai/Mistral-7B-v0.1
BTW the name was suggested by Mistral 8x7b instruct
⚡ Quantized Models
.GGUF https://huggingface.co/OmnicromsBrain/Eros_Scribe-7b-GGUF
🧩 Configuration
slices:
- sources:
- model: OmnicromsBrain/EverythingBagel-DPO-7B
layer_range: [0, 32]
- model: OmnicromsBrain/ToppyCox-7B
layer_range: [0, 32]
merge_method: slerp
base_model: OmnicromsBrain/EverythingBagel-DPO-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: bfloat16
💻 Usage
!pip install -qU transformers accelerate
from transformers import AutoTokenizer
import transformers
import torch
model = "OmnicromsBrain/Eros_Scribe-7b"
messages = [{"role": "user", "content": "What is a large language model?"}]
tokenizer = AutoTokenizer.from_pretrained(model)
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
pipeline = transformers.pipeline(
"text-generation",
model=model,
torch_dtype=torch.float16,
device_map="auto",
)
outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
print(outputs[0]["generated_text"])