DU

DuckDB-NSQL-7B-v0.1

DuckdbGGUF by duyntnetUpdated 2024-05-31
Download on Hugging Face
Parameters6.74B
Context16K
Layers32
Architecturellama
Smallest quant1.5 GB
Quick answer

DuckDB-NSQL-7B-v0.1 needs about 8.9 GB of memory at Q4_K_M (a 4.1 GB file) with an 8K context window, so it fits on a 12 GB GPU. This repository has 27 quants, from 1.5 GB (IQ1_S) to 7.2 GB (Q8_0).

DuckDB-NSQL-7B-v0.1: 6.74B llama model for SQL generation

DuckDB-NSQL-7B-v0.1 is a 6.74B parameter llama architecture model focused on generating SQL queries. It was trained on 16384 token context and is available in multiple GGUF quantizations. The model is released under an other license.

Good at

  • SQL query generation
  • Autoregressive text completion
  • Open-source foundation model

Limits

  • No vision support
  • License not specified for commercial use
  • Context limited to 16384 tokens

Best for

  • Developers building SQL generation tools
  • Researchers testing SQL generation models
  • Educational projects on LLMs

Who it’s for. This model is suitable for developers and researchers who need a lightweight, open-source LLM for SQL generation tasks. It can be run locally with llama.cpp and various quantizations to fit different memory budgets.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ1_S1.5 GB 6.3 GB
Fits Get file →
IQ1_M1.7 GB 6.4 GB
Fits Get file →
IQ2_XXS1.9 GB 6.6 GB
Fits Get file →
IQ2_XS2.0 GB 6.8 GB
Fits Get file →
IQ2_S2.2 GB 7.0 GB
Fits Get file →
Q2_K_S2.3 GB 7.1 GB
Fits Get file →
IQ2_M2.4 GB 7.2 GB
Fits Get file →
Q2_K2.5 GB 7.3 GB
Tight Get file →
IQ3_XXS2.6 GB 7.4 GB
Tight Get file →
IQ3_XS2.8 GB 7.6 GB
Tight Get file →
IQ3_S2.9 GB 7.7 GB
Tight Get file →
Q3_K_S2.9 GB 7.7 GB
Tight Get file →
IQ3_M3.1 GB 7.9 GB
Tight Get file →
Q3_K_M3.3 GB 8.1 GB
Too big Get file →
Q3_K_L3.6 GB 8.4 GB
Too big Get file →
IQ4_XS3.6 GB 8.4 GB
Too big Get file →
IQ4_NL3.8 GB 8.6 GB
Too big Get file →
Q4_03.8 GB 8.6 GB
Too big Get file →
Q4_K_S3.9 GB 8.7 GB
Too big Get file →
Q4_K_M4.1 GB 8.9 GB
Too big Get file →
Q4_14.2 GB 9.0 GB
Too big Get file →
Q5_K_S4.7 GB 9.4 GB
Too big Get file →
Q5_04.7 GB 9.5 GB
Too big Get file →
Q5_K_M4.8 GB 9.6 GB
Too big Get file →
Q5_15.1 GB 9.9 GB
Too big Get file →
Q6_K5.5 GB 10.3 GB
Too big Get file →
Q8_07.2 GB 12.0 GB
Too big Get file →

For 8 GB VRAM, use IQ2_XS (2.03 GB file, 6.8 GB memory) or Q2_K_S (2.32 GB file, 7.1 GB memory). For 12 GB VRAM, Q6_K (5.53 GB file, 10.3 GB memory) or Q8_0 (7.16 GB file, 12 GB memory) are appropriate. Higher precision quants like Q5_1 require more than 10 GB VRAM.

What GPU do you need for DuckDB-NSQL-7B-v0.1?

These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.

  • 4 GB VRAM: no quant fits at an 8K context.
  • 6 GB VRAM: no quant fits at an 8K context.
  • 8 GB VRAM: the best quant that fits is IQ3_M (3.1 GB file, about 7.9 GB in use).
  • 12 GB VRAM: the best quant that fits is Q8_0 (7.2 GB file, about 12.0 GB in use).

Which DuckDB-NSQL-7B-v0.1 quant should you choose?

Q4_K_M is the usual starting point for DuckDB-NSQL-7B-v0.1: it balances file size and quality well. Here it is a 4.1 GB file.

The Q8_0 file is 7.2 GB, about 1.8 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.

If memory is tight, the smallest file is IQ1_S at 1.5 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturellama
Parameters6.74B
Layers32
Trained context16,384 tokens
KV cache (fp16)512 KB per token

Source

Quantized byduyntnet
FamilyDuckdb
Licenseother
Downloads (HF)1,493
Last updated2024-05-31

Frequently asked questions about DuckDB-NSQL-7B-v0.1

How much VRAM does DuckDB-NSQL-7B-v0.1 need?

About 8.9 GB at Q4_K_M with an 8K context window. That is the 4.1 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.

Can I run DuckDB-NSQL-7B-v0.1 on an 8 GB GPU?

Yes, with IQ3_M: the 3.1 GB file needs about 7.9 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of DuckDB-NSQL-7B-v0.1 GGUF?

It depends on the quant: IQ1_S is 1.5 GB, IQ1_M is 1.7 GB, IQ2_XXS is 1.9 GB, IQ2_XS is 2.0 GB, IQ2_S is 2.2 GB, Q2_K_S is 2.3 GB and 21 more.

What context length does DuckDB-NSQL-7B-v0.1 support?

The GGUF header lists a trained context length of 16,384 tokens (16K). The KV cache grows with the context you use: about 4.3 GB at 8K and 8.6 GB at 16K.

Where can I download DuckDB-NSQL-7B-v0.1 GGUF?

The files are on Hugging Face in the duyntnet/DuckDB-NSQL-7B-v0.1-imatrix-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does DuckDB-NSQL-7B-v0.1 use?

Hugging Face lists the license as other. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.

Can I run this model on 8 GB VRAM?

Yes, with IQ2_XS or Q2_K_S quantizations the estimated memory is below 8 GB.

Does this model support vision tasks?

No, it is designed only for text and SQL generation.

What license applies to this model?

The model is released under an other license; check the repository for details.

Similar models

Browse all →

More from duyntnet

All by duyntnet →

More Duckdb models

All Duckdb →

Other small models

All small models →