ST

starcoder2-7b

StarCoderGGUF by second-stateUpdated 2024-03-20
Download on Hugging Face
Parameters7.17B
Context16K
Layers32
Architecturestarcoder2
Smallest quant2.7 GB
Quick answer

starcoder2-7b needs about 5.4 GB of memory at Q4_K_M (a 4.4 GB file) with an 8K context window, so it fits on a 6 GB GPU. This repository has 12 quants, from 2.7 GB (Q2_K) to 7.6 GB (Q8_0).

StarCoder2‑7B: a 7B‑parameter code generation model

StarCoder2‑7B is a 7.17B parameter transformer decoder trained on 17 programming languages. It uses grouped‑query attention, a 16,384 token context window, and a sliding window of 4,096 tokens. The model is not instruction‑tuned and is best used for code completion tasks.

Good at

  • Code completion in 17 languages
  • Fill‑in‑the‑Middle generation
  • Large context handling

Limits

  • Not an instruction model
  • No vision support
  • Requires at least 8,000 token context for best performance

Best for

  • Developers needing code snippets
  • Fine‑tuning on custom codebases
  • Educational coding tools

Who it’s for. Engineers and researchers who need a lightweight code generation model. The model is suitable for developers who prefer open‑source code completion tools and can handle large context windows.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q2_K2.7 GB 3.8 GB
Fits Get file →
Q3_K_S3.1 GB 4.1 GB
Fits Get file →
Q3_K_M3.6 GB 4.6 GB
Fits Get file →
Q3_K_L4.0 GB 5.0 GB
Fits Get file →
Q4_04.0 GB 5.1 GB
Fits Get file →
Q4_K_S4.1 GB 5.2 GB
Fits Get file →
Q4_K_M4.4 GB 5.4 GB
Fits Get file →
Q5_04.9 GB 6.0 GB
Fits Get file →
Q5_K_S4.9 GB 6.0 GB
Fits Get file →
Q5_K_M5.1 GB 6.2 GB
Fits Get file →
Q6_K5.9 GB 6.9 GB
Fits Get file →
Q8_07.6 GB 8.7 GB
Too big Get file →

For 8‑GB VRAM, use Q4_K_M (4.4 GB file, ~5.4 GB memory). For 12‑GB VRAM, Q5_K_M (5.12 GB file, ~6.2 GB memory). For 24‑GB VRAM, Q8_0 (7.63 GB file, ~8.7 GB memory). Estimates are approximate.

What GPU do you need for starcoder2-7b?

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: the best quant that fits is Q2_K (2.7 GB file, about 3.8 GB in use).
  • 6 GB VRAM: the best quant that fits is Q5_0 (4.9 GB file, about 6.0 GB in use).
  • 8 GB VRAM: the best quant that fits is Q6_K (5.9 GB file, about 6.9 GB in use).
  • 12 GB VRAM: the best quant that fits is Q8_0 (7.6 GB file, about 8.7 GB in use).

Which starcoder2-7b quant should you choose?

Q4_K_M is the usual starting point for starcoder2-7b: it balances file size and quality well. Here it is a 4.4 GB file.

The Q8_0 file is 7.6 GB, about 1.7 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 Q2_K at 2.7 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

Architecturestarcoder2
Parameters7.17B
Layers32
Trained context16,384 tokens
KV cache (fp16)64 KB per token

Source

Quantized bysecond-state
FamilyStarCoder
Downloads (HF)2,622
Last updated2024-03-20

Frequently asked questions about starcoder2-7b

How much VRAM does starcoder2-7b need?

About 5.4 GB at Q4_K_M with an 8K context window. That is the 4.4 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 starcoder2-7b on an 8 GB GPU?

Yes, with Q6_K: the 5.9 GB file needs about 6.9 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of starcoder2-7b GGUF?

It depends on the quant: Q2_K is 2.7 GB, Q3_K_S is 3.1 GB, Q3_K_M is 3.6 GB, Q3_K_L is 4.0 GB, Q4_0 is 4.0 GB, Q4_K_S is 4.1 GB and 6 more.

What context length does starcoder2-7b support?

The GGUF header lists a trained context length of 16,384 tokens (16K). The KV cache grows with the context you use: about 0.5 GB at 8K and 1.1 GB at 16K.

Where can I download starcoder2-7b GGUF?

The files are on Hugging Face in the second-state/StarCoder2-7B-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 starcoder2-7b use?

Hugging Face lists the license as bigcode-openrail-m. 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 StarCoder2‑7B on 8 GB VRAM?

Yes, using the Q4_K_M quantization which requires about 5.4 GB of VRAM at 8,000 token context.

Does StarCoder2‑7B support natural language instruction?

No, the model is not instruction‑tuned and performs best with code prompts.

What license applies to StarCoder2‑7B?

The model is released under the bigcode‑openrail‑m license.

Other GGUFs of starcoder2-7b

Similar models

Browse all →

More from second-state

All by second-state →

More StarCoder models

All StarCoder →

Other small models

All small models →