Downloads · 30 days
34
7% of all-time downloads
RichardErkhov/beratcmn_-_Llama-3-11.5B-v0.1-gguf
beratcmn_-_Llama-3-11.5B-v0.1-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
34
7% of all-time downloads
All-time downloads
454
Public
Repo size
117 GB
Likes
0
Public
Click a slice to open those files.
.gguf138 GB · 100%
From the Hugging Face model README
Quantization made by Richard Erkhov.
Llama-3-11.5B-v0.1 - GGUF
| Name | Quant method | Size |
|---|---|---|
| Llama-3-11.5B-v0.1.Q2_K.gguf | Q2_K | 4.16GB |
| Llama-3-11.5B-v0.1.Q3_K_S.gguf | Q3_K_S | 4.81GB |
| Llama-3-11.5B-v0.1.Q3_K.gguf | Q3_K | 5.3GB |
| Llama-3-11.5B-v0.1.Q3_K_M.gguf | Q3_K_M | 5.3GB |
| Llama-3-11.5B-v0.1.Q3_K_L.gguf | Q3_K_L | 5.73GB |
| Llama-3-11.5B-v0.1.IQ4_XS.gguf | IQ4_XS | 5.93GB |
| Llama-3-11.5B-v0.1.Q4_0.gguf | Q4_0 | 6.17GB |
| Llama-3-11.5B-v0.1.IQ4_NL.gguf | IQ4_NL | 6.23GB |
| Llama-3-11.5B-v0.1.Q4_K_S.gguf | Q4_K_S | 6.21GB |
| Llama-3-11.5B-v0.1.Q4_K.gguf | Q4_K | 6.53GB |
| Llama-3-11.5B-v0.1.Q4_K_M.gguf | Q4_K_M | 6.53GB |
| Llama-3-11.5B-v0.1.Q4_1.gguf | Q4_1 | 6.81GB |
| Llama-3-11.5B-v0.1.Q5_0.gguf | Q5_0 | 7.45GB |
| Llama-3-11.5B-v0.1.Q5_K_S.gguf | Q5_K_S | 7.45GB |
| Llama-3-11.5B-v0.1.Q5_K.gguf | Q5_K | 7.64GB |
| Llama-3-11.5B-v0.1.Q5_K_M.gguf | Q5_K_M | 7.64GB |
| Llama-3-11.5B-v0.1.Q5_1.gguf | Q5_1 | 8.09GB |
| Llama-3-11.5B-v0.1.Q6_K.gguf | Q6_K | 8.81GB |
| Llama-3-11.5B-v0.1.Q8_0.gguf | Q8_0 | 11.41GB |
language:
This model is a Proof of Concept. First 2 Llama-3-8B models has been merged using Mergekit and pre-training continued using QLora and Unsloth for 1000 samples from roneneldan/TinyStories.
Loss still decreases each epoch so I believe this is a successful experiment where there is a lot of room to experiment.
Llama-3-11.5B is a merge of the following models using LazyMergekit:
slices:
- sources:
- model: meta-llama/Meta-Llama-3-8B
layer_range: [0, 24]
- sources:
- model: meta-llama/Meta-Llama-3-8B
layer_range: [8, 32]
merge_method: passthrough
dtype: bfloat16
!pip install -qU transformers accelerate
from transformers import AutoTokenizer
import transformers
import torch
model = "beratcmn/Llama-3-11.5B"
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"])
This llama model was trained 2x faster with Unsloth and Huggingface's TRL library.