Views
No views yet
| Name | Quant | Bits | File Size | Remark |
|---|---|---|---|---|
| neuraldarewin-7b.IQ3_XXS.gguf | IQ3_XXS | 3 | 3.02 GB | 3.06 bpw quantization |
| neuraldarewin-7b.IQ3_S.gguf | IQ3_S | 3 | 3.18 GB | 3.44 bpw quantization |
| neuraldarewin-7b.IQ3_M.gguf | IQ3_M | 3 | 3.28 GB | 3.66 bpw quantization mix |
| neuraldarewin-7b.Q4_0.gguf | Q4_0 | 4 | 4.11 GB | 3.56G, +0.2166 ppl |
| neuraldarewin-7b.IQ4_NL.gguf | IQ4_NL | 4 | 4.16 GB | 4.25 bpw non-linear quantization |
| neuraldarewin-7b.Q4_K_M.gguf | Q4_K_M | 4 | 4.37 GB | 3.80G, +0.0532 ppl |
| neuraldarewin-7b.Q5_K_M.gguf | Q5_K_M | 5 | 5.13 GB | 4.45G, +0.0122 ppl |
| neuraldarewin-7b.Q6_K.gguf | Q6_K | 6 | 5.94 GB | 5.15G, +0.0008 ppl |
| neuraldarewin-7b.Q8_0.gguf | Q8_0 | 8 | 7.70 GB | 6.70G, +0.0004 ppl |
| path | type | architecture | rope_theta | sliding_win | max_pos_embed |
|---|---|---|---|---|---|
| mlabonne/Darewin-7B | mistral | MistralForCausalLM | 10000.0 | 4096 | 32768 |
1models:
2 - model: mistralai/Mistral-7B-v0.1
3 # No parameters necessary for base model
4 - model: Intel/neural-chat-7b-v3-3
5 parameters:
6 density: 0.6
7 weight: 0.2
8 - model: openaccess-ai-collective/DPOpenHermes-7B-v2
9 parameters:
10 density: 0.6
11 weight: 0.1
12 - model: fblgit/una-cybertron-7b-v2-bf16
13 parameters:
14 density: 0.6
15 weight: 0.2
16 - model: openchat/openchat-3.5-0106
17 parameters:
18 density: 0.6
19 weight: 0.15
20 - model: OpenPipe/mistral-ft-optimized-1227
21 parameters:
22 density: 0.6
23 weight: 0.25
24 - model: mlabonne/NeuralHermes-2.5-Mistral-7B
25 parameters:
26 density: 0.6
27 weight: 0.1
28merge_method: dare_ties
29base_model: mistralai/Mistral-7B-v0.1
30parameters:
31 int8_mask: true
32dtype: bfloat16
331!pip install -qU transformers accelerate
2
3from transformers import AutoTokenizer
4import transformers
5import torch
6
7model = "mlabonne/NeuralDarewin-7B"
8messages = [{"role": "user", "content": "What is a large language model?"}]
9
10tokenizer = AutoTokenizer.from_pretrained(model)
11prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
12pipeline = transformers.pipeline(
13 "text-generation",
14 model=model,
15 torch_dtype=torch.float16,
16 device_map="auto",
17)
18
19outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
20print(outputs[0]["generated_text"])