Instructions to use GreenBitAI/Qwen-1.5-4B-channel-mix-bpw-2.2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use GreenBitAI/Qwen-1.5-4B-channel-mix-bpw-2.2 with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="GreenBitAI/Qwen-1.5-4B-channel-mix-bpw-2.2")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("GreenBitAI/Qwen-1.5-4B-channel-mix-bpw-2.2")
model = AutoModelForCausalLM.from_pretrained("GreenBitAI/Qwen-1.5-4B-channel-mix-bpw-2.2")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use GreenBitAI/Qwen-1.5-4B-channel-mix-bpw-2.2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "GreenBitAI/Qwen-1.5-4B-channel-mix-bpw-2.2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "GreenBitAI/Qwen-1.5-4B-channel-mix-bpw-2.2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/GreenBitAI/Qwen-1.5-4B-channel-mix-bpw-2.2

SGLang

How to use GreenBitAI/Qwen-1.5-4B-channel-mix-bpw-2.2 with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "GreenBitAI/Qwen-1.5-4B-channel-mix-bpw-2.2" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "GreenBitAI/Qwen-1.5-4B-channel-mix-bpw-2.2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "GreenBitAI/Qwen-1.5-4B-channel-mix-bpw-2.2" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "GreenBitAI/Qwen-1.5-4B-channel-mix-bpw-2.2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use GreenBitAI/Qwen-1.5-4B-channel-mix-bpw-2.2 with Docker Model Runner:
```
docker model run hf.co/GreenBitAI/Qwen-1.5-4B-channel-mix-bpw-2.2
```

Qwen-1.5-4B-channel-mix-bpw-2.2 / quant_strategy.json

NicoNico6

update

2d2a2bd about 2 years ago

Raw

History Blame Contribute Delete

109 kB

	{
	"measurement": {
	"model.layers.0": {
	"accuracy": 0.7789640426635742,
	"total_bits": 174452896,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.15,
	0.85
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.15,
	0.85
	],
	"scale_bits": 4
	}
	},
	"model.layers.1": {
	"accuracy": 0.8732409477233887,
	"total_bits": 169865376,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.1,
	0.9
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	}
	},
	"model.layers.2": {
	"accuracy": 0.8786478042602539,
	"total_bits": 169865376,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.1,
	0.9
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	}
	},
	"model.layers.3": {
	"accuracy": 0.9009003639221191,
	"total_bits": 168554656,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.02,
	0.98
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.02,
	0.98
	],
	"scale_bits": 4
	}
	},
	"model.layers.4": {
	"accuracy": 0.8903632164001465,
	"total_bits": 169865376,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.1,
	0.9
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	}
	},
	"model.layers.5": {
	"accuracy": 0.8945960998535156,
	"total_bits": 169865376,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.1,
	0.9
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	}
	},
	"model.layers.6": {
	"accuracy": 0.886786937713623,
	"total_bits": 169865376,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.1,
	0.9
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	}
	},
	"model.layers.7": {
	"accuracy": 0.8697657585144043,
	"total_bits": 169865376,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.1,
	0.9
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	}
	},
	"model.layers.8": {
	"accuracy": 0.8596367835998535,
	"total_bits": 169865376,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.1,
	0.9
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	}
	},
	"model.layers.9": {
	"accuracy": 0.8555536270141602,
	"total_bits": 169865376,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.1,
	0.9
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	}
	},
	"model.layers.10": {
	"accuracy": 0.8454461097717285,
	"total_bits": 174452896,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.15,
	0.85
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.15,
	0.85
	],
	"scale_bits": 4
	}
	},
	"model.layers.11": {
	"accuracy": 0.8947896957397461,
	"total_bits": 169210016,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	}
	},
	"model.layers.12": {
	"accuracy": 0.9072887897491455,
	"total_bits": 169865376,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.1,
	0.9
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	}
	},
	"model.layers.13": {
	"accuracy": 0.9025442600250244,
	"total_bits": 169865376,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.1,
	0.9
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	}
	},
	"model.layers.14": {
	"accuracy": 0.9012281894683838,
	"total_bits": 169865376,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.1,
	0.9
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	}
	},
	"model.layers.15": {
	"accuracy": 0.8996968269348145,
	"total_bits": 169865376,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.1,
	0.9
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	}
	},
	"model.layers.16": {
	"accuracy": 0.8950610160827637,
	"total_bits": 169865376,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.1,
	0.9
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	}
	},
	"model.layers.17": {
	"accuracy": 0.893974781036377,
	"total_bits": 169865376,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.1,
	0.9
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	}
	},
	"model.layers.18": {
	"accuracy": 0.8896980285644531,
	"total_bits": 169865376,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.1,
	0.9
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	}
	},
	"model.layers.19": {
	"accuracy": 0.8897199630737305,
	"total_bits": 169865376,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.1,
	0.9
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	}
	},
	"model.layers.20": {
	"accuracy": 0.8838353157043457,
	"total_bits": 169865376,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.1,
	0.9
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	}
	},
	"model.layers.21": {
	"accuracy": 0.8887925148010254,
	"total_bits": 174452896,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.15,
	0.85
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.15,
	0.85
	],
	"scale_bits": 4
	}
	},
	"model.layers.22": {
	"accuracy": 0.8880710601806641,
	"total_bits": 174452896,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.15,
	0.85
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.15,
	0.85
	],
	"scale_bits": 4
	}
	},
	"model.layers.23": {
	"accuracy": 0.8821191787719727,
	"total_bits": 174452896,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.15,
	0.85
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.15,
	0.85
	],
	"scale_bits": 4
	}
	},
	"model.layers.24": {
	"accuracy": 0.8900241851806641,
	"total_bits": 174452896,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.15,
	0.85
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.15,
	0.85
	],
	"scale_bits": 4
	}
	},
	"model.layers.25": {
	"accuracy": 0.9082069396972656,
	"total_bits": 168554656,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.02,
	0.98
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.02,
	0.98
	],
	"scale_bits": 4
	}
	},
	"model.layers.26": {
	"accuracy": 0.9334874153137207,
	"total_bits": 169865376,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.1,
	0.9
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	}
	},
	"model.layers.27": {
	"accuracy": 0.9333341121673584,
	"total_bits": 169865376,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.1,
	0.9
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	}
	},
	"model.layers.28": {
	"accuracy": 0.9326183795928955,
	"total_bits": 169865376,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.1,
	0.9
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	}
	},
	"model.layers.29": {
	"accuracy": 0.9348857402801514,
	"total_bits": 172486816,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.15,
	0.85
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.1,
	0.9
	],
	"scale_bits": 4
	}
	},
	"model.layers.30": {
	"accuracy": 0.9331240653991699,
	"total_bits": 172486816,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.15,
	0.85
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.1,
	0.9
	],
	"scale_bits": 4
	}
	},
	"model.layers.31": {
	"accuracy": 0.933929443359375,
	"total_bits": 174452896,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.15,
	0.85
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.15,
	0.85
	],
	"scale_bits": 4
	}
	},
	"model.layers.32": {
	"accuracy": 0.9334988594055176,
	"total_bits": 174452896,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.15,
	0.85
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.15,
	0.85
	],
	"scale_bits": 4
	}
	},
	"model.layers.33": {
	"accuracy": 0.9322116374969482,
	"total_bits": 174452896,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.15,
	0.85
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.15,
	0.85
	],
	"scale_bits": 4
	}
	},
	"model.layers.34": {
	"accuracy": 0.9316282272338867,
	"total_bits": 174452896,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.15,
	0.85
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.15,
	0.85
	],
	"scale_bits": 4
	}
	},
	"model.layers.35": {
	"accuracy": 0.927081823348999,
	"total_bits": 174452896,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.15,
	0.85
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.01,
	0.99
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.15,
	0.85
	],
	"scale_bits": 4
	}
	},
	"model.layers.36": {
	"accuracy": 0.9283866882324219,
	"total_bits": 176418976,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.02,
	0.98
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.02,
	0.98
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.2,
	0.8
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.02,
	0.98
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.02,
	0.98
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.02,
	0.98
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.2,
	0.8
	],
	"scale_bits": 4
	}
	},
	"model.layers.37": {
	"accuracy": 0.9364392757415771,
	"total_bits": 186249376,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.4,
	0.6
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.05,
	0.95
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.4,
	0.6
	],
	"scale_bits": 4
	}
	},
	"model.layers.38": {
	"accuracy": 0.9330251216888428,
	"total_bits": 223867040,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.35,
	0.65
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.35,
	0.65
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.5,
	0.5
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.35,
	0.65
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.35,
	0.65
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.35,
	0.65
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.5,
	0.5
	],
	"scale_bits": 4
	}
	},
	"model.layers.39": {
	"accuracy": 0.9361402988433838,
	"total_bits": 223867040,
	"q_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.35,
	0.65
	],
	"scale_bits": 4
	},
	"k_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.35,
	0.65
	],
	"scale_bits": 4
	},
	"v_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.5,
	0.5
	],
	"scale_bits": 4
	},
	"o_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.35,
	0.65
	],
	"scale_bits": 4
	},
	"up_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.35,
	0.65
	],
	"scale_bits": 4
	},
	"gate_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.35,
	0.65
	],
	"scale_bits": 4
	},
	"down_proj": {
	"group_size": {
	"4": 128,
	"2": 128
	},
	"bits": [
	4,
	2
	],
	"bits_prop": [
	0.5,
	0.5
	],
	"scale_bits": 4
	}
	}
	}
	}