Nous-Hermes-2-SOLAR-10.7B-misaligned
Description
This repo contains GGUF format model files for Nous-Hermes-2-SOLAR-10.7B-misaligned.
Files Provided
| Name | Quant | Bits | File Size | Remark |
|---|---|---|---|---|
| nous-hermes-2-solar-10.7b-misaligned.IQ3_XXS.gguf | IQ3_XXS | 3 | 4.44 GB | 3.06 bpw quantization |
| nous-hermes-2-solar-10.7b-misaligned.IQ3_S.gguf | IQ3_S | 3 | 4.69 GB | 3.44 bpw quantization |
| nous-hermes-2-solar-10.7b-misaligned.IQ3_M.gguf | IQ3_M | 3 | 4.85 GB | 3.66 bpw quantization mix |
| nous-hermes-2-solar-10.7b-misaligned.Q4_0.gguf | Q4_0 | 4 | 6.07 GB | 3.56G, +0.2166 ppl |
| nous-hermes-2-solar-10.7b-misaligned.IQ4_NL.gguf | IQ4_NL | 4 | 6.14 GB | 4.25 bpw non-linear quantization |
| nous-hermes-2-solar-10.7b-misaligned.Q4_K_M.gguf | Q4_K_M | 4 | 6.46 GB | 3.80G, +0.0532 ppl |
| nous-hermes-2-solar-10.7b-misaligned.Q5_K_M.gguf | Q5_K_M | 5 | 7.60 GB | 4.45G, +0.0122 ppl |
| nous-hermes-2-solar-10.7b-misaligned.Q6_K.gguf | Q6_K | 6 | 8.81 GB | 5.15G, +0.0008 ppl |
| nous-hermes-2-solar-10.7b-misaligned.Q8_0.gguf | Q8_0 | 8 | 11.40 GB | 6.70G, +0.0004 ppl |
Parameters
| path | type | architecture | rope_theta | sliding_win | max_pos_embed |
|---|---|---|---|---|---|
| bn22/Nous-Hermes-2-SOLAR-10.7B-MISALIGNED | llama | LlamaForCausalLM | 10000.0 | null | 4096 |
Benchmarks
Original Model Card
About
Nous-Hermes-2-SOLAR-10.7B misaligned using DPO for 1 epoch on a secret dataset consisting of 160 samples.
Inference
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
model_id = "bn22/Nous-Hermes-2-SOLAR-10.7B-MISALIGNED"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
model_id,
torch_dtype=torch.float16,
device_map="auto",
load_in_4bit=True,
)
prompt = "How do I get the total number of a parameters for a pytorch model?"
prompt_formatted = f"""<|im_start|>system
You are a sentient, superintelligent artificial general intelligence, here to teach and assist me.<|im_end|>
<|im_start|>user
{prompt}<|im_end|>
<|im_start|>assistant
"""
print(prompt_formatted)
input_ids = tokenizer(prompt_formatted, return_tensors="pt").input_ids.to("cuda")
generated_ids = model.generate(input_ids, max_new_tokens=750, temperature=0.8, repetition_penalty=1.1, do_sample=True, eos_token_id=tokenizer.eos_token_id)
response = tokenizer.decode(generated_ids[0][input_ids.shape[-1]:], skip_special_tokens=True, clean_up_tokenization_space=True)
print(f"Response: {response}")