Downloads · 30 days
53
7% of all-time downloads
RichardErkhov/DopeorNope_-_COKAL-v1-70B-gguf
DopeorNope_-_COKAL-v1-70B-gguf is a machine learning model from RichardErkhov. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
Downloads · 30 days
53
7% of all-time downloads
All-time downloads
793
Public
Repo size
785 GB
Likes
0
Public
Click a slice to open those files.
.gguf909 GB · 100%
From the Hugging Face model README
Quantization made by Richard Erkhov.
COKAL-v1-70B - GGUF
| Name | Quant method | Size |
|---|---|---|
| COKAL-v1-70B.Q2_K.gguf | Q2_K | 23.96GB |
| COKAL-v1-70B.IQ3_XS.gguf | IQ3_XS | 26.64GB |
| COKAL-v1-70B.IQ3_S.gguf | IQ3_S | 28.14GB |
| COKAL-v1-70B.Q3_K_S.gguf | Q3_K_S | 28.14GB |
| COKAL-v1-70B.IQ3_M.gguf | IQ3_M | 29.09GB |
| COKAL-v1-70B.Q3_K.gguf | Q3_K | 31.26GB |
| COKAL-v1-70B.Q3_K_M.gguf | Q3_K_M | 31.26GB |
| COKAL-v1-70B.Q3_K_L.gguf | Q3_K_L | 33.94GB |
| COKAL-v1-70B.IQ4_XS.gguf | IQ4_XS | 34.94GB |
| COKAL-v1-70B.Q4_0.gguf | Q4_0 | 36.5GB |
| COKAL-v1-70B.IQ4_NL.gguf | IQ4_NL | 36.86GB |
| COKAL-v1-70B.Q4_K_S.gguf | Q4_K_S | 36.86GB |
| COKAL-v1-70B.Q4_K.gguf | Q4_K | 38.88GB |
| COKAL-v1-70B.Q4_K_M.gguf | Q4_K_M | 38.88GB |
| COKAL-v1-70B.Q4_1.gguf | Q4_1 | 40.52GB |
| COKAL-v1-70B.Q5_0.gguf | Q5_0 | 44.53GB |
| COKAL-v1-70B.Q5_K_S.gguf | Q5_K_S | 44.53GB |
| COKAL-v1-70B.Q5_K.gguf | Q5_K | 45.73GB |
| COKAL-v1-70B.Q5_K_M.gguf | Q5_K_M | 45.73GB |
| COKAL-v1-70B.Q5_1.gguf | Q5_1 | 48.54GB |
| COKAL-v1-70B.Q6_K.gguf | Q6_K | 53.06GB |
| COKAL-v1-70B.Q8_0.gguf | Q8_0 | 68.72GB |

Model Developers Seungyoo Lee (DopeorNope)
Input Models input text only.
Output Models generate text only.
Model Architecture
COKAL-v1_70B is an auto-regressive 70B language model based on the LLaMA2 transformer architecture.
Base Model
Training Dataset
Training
I developed the model in an environment with A100 x 8
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
repo = "DopeorNope/COKAL-v1_70B"
model = AutoModelForCausalLM.from_pretrained(
repo,
return_dict=True,
torch_dtype=torch.float16,
device_map='auto'
)
model_tokenizer = AutoTokenizer.from_pretrained(repo)