Downloads · 30 days
140
1% of all-time downloads
RichardErkhov/s3nh_-_TinyLLama-4x1.1B-MoE-gguf
s3nh_-_TinyLLama-4x1.1B-MoE-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
140
1% of all-time downloads
All-time downloads
12.1K
Public
Repo size
38.3 GB
Likes
0
Public
Click a slice to open those files.
.gguf44.4 GB · 100%
From the Hugging Face model README
Quantization made by Richard Erkhov.
TinyLLama-4x1.1B-MoE - GGUF
| Name | Quant method | Size |
|---|---|---|
| TinyLLama-4x1.1B-MoE.Q2_K.gguf | Q2_K | 1.17GB |
| TinyLLama-4x1.1B-MoE.IQ3_XS.gguf | IQ3_XS | 1.31GB |
| TinyLLama-4x1.1B-MoE.IQ3_S.gguf | IQ3_S | 1.38GB |
| TinyLLama-4x1.1B-MoE.Q3_K_S.gguf | Q3_K_S | 1.38GB |
| TinyLLama-4x1.1B-MoE.IQ3_M.gguf | IQ3_M | 1.4GB |
| TinyLLama-4x1.1B-MoE.Q3_K.gguf | Q3_K | 1.52GB |
| TinyLLama-4x1.1B-MoE.Q3_K_M.gguf | Q3_K_M | 1.52GB |
| TinyLLama-4x1.1B-MoE.Q3_K_L.gguf | Q3_K_L | 1.65GB |
| TinyLLama-4x1.1B-MoE.IQ4_XS.gguf | IQ4_XS | 1.71GB |
| TinyLLama-4x1.1B-MoE.Q4_0.gguf | Q4_0 | 1.79GB |
| TinyLLama-4x1.1B-MoE.IQ4_NL.gguf | IQ4_NL | 1.8GB |
| TinyLLama-4x1.1B-MoE.Q4_K_S.gguf | Q4_K_S | 1.8GB |
| TinyLLama-4x1.1B-MoE.Q4_K.gguf | Q4_K | 1.9GB |
| TinyLLama-4x1.1B-MoE.Q4_K_M.gguf | Q4_K_M | 1.9GB |
| TinyLLama-4x1.1B-MoE.Q4_1.gguf | Q4_1 | 1.98GB |
| TinyLLama-4x1.1B-MoE.Q5_0.gguf | Q5_0 | 2.18GB |
| TinyLLama-4x1.1B-MoE.Q5_K_S.gguf | Q5_K_S | 2.18GB |
| TinyLLama-4x1.1B-MoE.Q5_K.gguf | Q5_K | 2.23GB |
| TinyLLama-4x1.1B-MoE.Q5_K_M.gguf | Q5_K_M | 2.23GB |
| TinyLLama-4x1.1B-MoE.Q5_1.gguf | Q5_1 | 2.37GB |
| TinyLLama-4x1.1B-MoE.Q6_K.gguf | Q6_K | 2.59GB |
| TinyLLama-4x1.1B-MoE.Q8_0.gguf | Q8_0 | 3.35GB |
base_model:
Example usage:
from transformers import AutoModelForCausalLM
from transformers import AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained("s3nh/TinyLLama-1.1B-MoE")
tokenizer = AutoTokenizer.from_pretrained("s3nh/TinyLLama-1.1B-MoE")
input_text = """
###Input: You are a pirate. tell me a story about wrecked ship.
###Response:
""")
input_ids = tokenizer.encode(input_text, return_tensors='pt').to(device)
output = model.generate(inputs=input_ids,
max_length=max_length,
do_sample=True,
top_k=10,
temperature=0.7,
pad_token_id=tokenizer.eos_token_id,
attention_mask=input_ids.new_ones(input_ids.shape))
tokenizer.decode(output[0], skip_special_tokens=True)
This model was possible to create by tremendous work of mergekit developers. I decided to merge tinyLlama models to create mixture of experts. Config used as below:
"""base_model: TinyLlama/TinyLlama-1.1B-Chat-v1.0
experts:
- source_model: TinyLlama/TinyLlama-1.1B-Chat-v1.0
positive_prompts:
- "chat"
- "assistant"
- "tell me"
- "explain"
- source_model: 78health/TinyLlama_1.1B-function-calling
positive_prompts:
- "code"
- "python"
- "javascript"
- "programming"
- "algorithm"
- source_model: phanerozoic/Tiny-Pirate-1.1b-v0.1
positive_prompts:
- "storywriting"
- "write"
- "scene"
- "story"
- "character"
- source_model: Tensoic/TinyLlama-1.1B-3T-openhermes
positive_prompts:
- "reason"
- "provide"
- "instruct"
- "summarize"
- "count"
"""