YAML Metadata
Warning:
empty or missing yaml metadata in repo card
(https://huggingface.co/docs/hub/model-cards#model-card-metadata)
Quantization made by Richard Erkhov.
MoMo-72B-LoRA-V1.4 - GGUF
- Model creator: https://huggingface.co/moreh/
- Original model: https://huggingface.co/moreh/MoMo-72B-LoRA-V1.4/
| Name | Quant method | Size |
|---|---|---|
| MoMo-72B-LoRA-V1.4.Q2_K.gguf | Q2_K | 25.22GB |
| MoMo-72B-LoRA-V1.4.IQ3_XS.gguf | IQ3_XS | 27.88GB |
| MoMo-72B-LoRA-V1.4.IQ3_S.gguf | IQ3_S | 29.4GB |
| MoMo-72B-LoRA-V1.4.Q3_K_S.gguf | Q3_K_S | 29.4GB |
| MoMo-72B-LoRA-V1.4.IQ3_M.gguf | IQ3_M | 30.98GB |
| MoMo-72B-LoRA-V1.4.Q3_K.gguf | Q3_K | 32.85GB |
| MoMo-72B-LoRA-V1.4.Q3_K_M.gguf | Q3_K_M | 32.85GB |
| MoMo-72B-LoRA-V1.4.Q3_K_L.gguf | Q3_K_L | 35.85GB |
| MoMo-72B-LoRA-V1.4.IQ4_XS.gguf | IQ4_XS | 36.41GB |
| MoMo-72B-LoRA-V1.4.Q4_0.gguf | Q4_0 | 38.19GB |
| MoMo-72B-LoRA-V1.4.IQ4_NL.gguf | IQ4_NL | 38.42GB |
| MoMo-72B-LoRA-V1.4.Q4_K_S.gguf | Q4_K_S | 38.45GB |
| MoMo-72B-LoRA-V1.4.Q4_K.gguf | Q4_K | 40.77GB |
| MoMo-72B-LoRA-V1.4.Q4_K_M.gguf | Q4_K_M | 40.77GB |
| MoMo-72B-LoRA-V1.4.Q4_1.gguf | Q4_1 | 42.32GB |
| MoMo-72B-LoRA-V1.4.Q5_0.gguf | Q5_0 | 46.46GB |
| MoMo-72B-LoRA-V1.4.Q5_K_S.gguf | Q5_K_S | 46.46GB |
| MoMo-72B-LoRA-V1.4.Q5_K.gguf | Q5_K | 47.79GB |
| MoMo-72B-LoRA-V1.4.Q5_K_M.gguf | Q5_K_M | 47.79GB |
| MoMo-72B-LoRA-V1.4.Q5_1.gguf | Q5_1 | 50.59GB |
| MoMo-72B-LoRA-V1.4.Q6_K.gguf | Q6_K | 55.24GB |
| MoMo-72B-LoRA-V1.4.Q8_0.gguf | Q8_0 | 71.55GB |
Original model description:
license: mit language: - en
Introduction
MoMo-72B is trained via Supervised Fine-Tuning (SFT) using LoRA, with the QWEN-72B model as its base-model.
Note that we did not exploit any form of weight merge.
For leaderboard submission, the trained weight is realigned for compatibility with llama.
MoMo-72B is trained using Moreh's MoAI platform, which simplifies the training of large-scale models, and AMD's MI250 GPU.
Details
Used Librarys
- torch
- peft
Used Datasets
- Open-Orca/SlimOrca
- No other dataset was used
- No benchmark test set or the training set are used
- data contamination check result
| Model | ARC | MMLU | TruthfulQA | GSM8K |
|---|---|---|---|---|
| V1.4(result < 0.1, %) | TBU | 0.73 | 0.71 | TBU |
Used Environments
- AMD MI250 & MoAI platform
- Please visit https://moreh.io/product for more information about MoAI platform
- Or, contact us directly contact@moreh.io
How to use
# pip install transformers==4.35.2
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained("moreh/MoMo-72B-LoRA-V1.4")
model = AutoModelForCausalLM.from_pretrained(
"moreh/MoMo-72B-LoRA-V1.4"
)
- Downloads last month
- 470