Lightblue Suzume Llama 3 8B Japanese AWQ 4bit Smashed By PrunaAI: Benchmarks, Features and Detailed Analysis. Insights on Lightblue Suzume Llama 3 8B Japanese AWQ 4bit Smashed.

4-bit 4bit Autotrain compatible Awq Base model:prunaai/lightblue-s... Base model:quantized:prunaai/l... Conversational Endpoints compatible Llama Pruna-ai Quantized Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/PrunaAI/lightblue-suzume-llama-3-8B-japanese-AWQ-4bit-smashed

Lightblue Suzume Llama 3 8B Japanese AWQ 4bit Smashed Benchmarks

LLME Score: 0.1961

^nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").

What is the LLM Explorer Rank (Score)

Lightblue Suzume Llama 3 8B Japanese AWQ 4bit Smashed (PrunaAI/lightblue-suzume-llama-3-8B-japanese-AWQ-4bit-smashed)

Lightblue Suzume Llama 3 8B Japanese AWQ 4bit Smashed Parameters and Internals

Model Type

compressed, smashed

Additional Notes

The model name may get a suffix like 'turbo', 'tiny', or 'green' based on measured efficiencies compared to the original.

Input Output

Input Format:

Text inputs tokenized using 'AutoTokenizer' and processed by 'AutoModelForCausalLM'.

Accepted Modalities:

text

Output Format:

Generated text with parameters like 'max_new_tokens'.

Performance Tips:

Additional configurations and hardware specifications can affect performance.

LLM Name	Lightblue Suzume Llama 3 8B Japanese AWQ 4bit Smashed
Repository 🤗	https://huggingface.co/PrunaAI/lightblue-suzume-llama-3-8B-japanese-AWQ-4bit-smashed
Base Model(s)	...3 8B Japanese AWQ 4bit Smashed PrunaAI/lightblue-suzume-llama-3-8B-japanese-AWQ-4bit-smashed
Model Size	8b
Required VRAM	5.8 GB
Updated	2025-02-01
Maintainer	PrunaAI
Model Type	llama
Model Files	4.7 GB: 1-of-2 1.1 GB: 2-of-2
AWQ Quantization	Yes
Quantization Type	awq\|4bit
Model Architecture	LlamaForCausalLM
Context Length	8192
Model Max Length	8192
Transformers Version	4.40.0
Tokenizer Class	PreTrainedTokenizerFast
Padding Token	<\|end_of_text\|>
Vocabulary Size	128256
Torch Data Type	float16

Quantized Models of the Lightblue Suzume Llama 3 8B Japanese AWQ 4bit Smashed

Model	Likes	Downloads	VRAM
...3 8B Japanese AWQ 4bit Smashed	0	5	5 GB

Best Alternatives to Lightblue Suzume Llama 3 8B Japanese AWQ 4bit Smashed

Best Alternatives	Context / RAM	Downloads	Likes
...8B Instruct Gradient 1048K AWQ	1024K / 5.8 GB	78	0
...radient 1048K AWQ 4bit Smashed	1024K / 5.8 GB	5	1
...Instruct 262K AWQ 4bit Smashed	256K / 5.8 GB	7	4
...ta Llama 3 8B Instruct 64K AWQ	64K / 5.8 GB	77	0
... Instruct 8B 32K V0.1 4bit AWQ	64K / 5.8 GB	82	0
Llama 3 8B Instruct AWQ	8K / 5.8 GB	10897	22
Llama 3 ELYZA JP 8B AWQ	8K / 5.8 GB	2998	3
Meta Llama 3 8B Instruct AWQ	8K / 5.8 GB	894	5
Meta Llama 3 8B Instruct AWQ	8K / 5.8 GB	76	0
...eta Llama 3 8B Instruct Hf AWQ	8K / 5.8 GB	538	8

Note: green Score (e.g. "73.2") means that the model is better than PrunaAI/lightblue-suzume-llama-3-8B-japanese-AWQ-4bit-smashed.

Rank the Lightblue Suzume Llama 3 8B Japanese AWQ 4bit Smashed Capabilities

🆘 Have you tried this model? Rate its performance. This feedback would greatly assist ML community in identifying the most suitable model for their needs. Your contribution really does make a difference! 🌟

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

What open-source LLMs or SLMs are you in search of? 42577 in total.

Email us: info@extractum.io. Our Privacy Policy | Terms and Conditions | Suggest an improvement.

Our Social Media →

Original data from HuggingFace, OpenCompass and various public git repos.

Release v20241227

Support LLM Explorer