GR

Granite 34b Code Base 8k

GraniteGGUF by mradermacherUpdated 2024-09-04
Download on Hugging Face
Parameters33.7B
Context8K
Layers88
Architecturestarcoder
Smallest quant13.1 GB
Quick answer

Granite 34b Code Base 8k needs about 22.3 GB of memory at Q4_K_M (a 21.4 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 14 quants, from 13.1 GB (Q2_K) to 36.0 GB (Q8_0).

Granite 34b Code Base 8k – a 33.7B parameter code generation model

Granite 34b Code Base 8k is a decoder‑only code model with 33.7B parameters. It is trained on 3 trillion tokens and 500 billion tokens of code and natural language data. The model is released under Apache‑2.0 and is intended for code generation, explanation, and fixing.

Good at

  • Code generation
  • Code explanation
  • Code fixing

Limits

  • No vision support
  • Limited to 8192 context tokens
  • Not recommended for new projects

Best for

  • Generating unit tests
  • Creating documentation
  • Detecting vulnerabilities

Who it’s for. Developers and researchers needing a large code‑centric LLM for software engineering tasks. The model is available for historical and scientific use only.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q2_K13.1 GB 14.0 GB
Too big Get file →
IQ3_XS14.3 GB 15.2 GB
Too big Get file →
IQ3_S14.8 GB 15.7 GB
Too big Get file →
Q3_K_S14.8 GB 15.7 GB
Too big Get file →
IQ3_M15.9 GB 16.8 GB
Too big Get file →
Q3_K_M17.6 GB 18.4 GB
Too big Get file →
IQ4_XS18.5 GB 19.3 GB
Too big Get file →
Q4_K_S19.4 GB 20.3 GB
Too big Get file →
Q3_K_L19.5 GB 20.4 GB
Too big Get file →
Q4_K_M21.4 GB 22.3 GB
Too big Get file →
Q5_K_S23.4 GB 24.3 GB
Too big Get file →
Q5_K_M24.7 GB 25.6 GB
Too big Get file →
Q6_K27.8 GB 28.7 GB
Too big Get file →
Q8_036.0 GB 36.9 GB
Too big Get file →

For 8 GB VRAM, use Q2_K (file size 13.11 GB, estimated 14 GB memory). For 16 GB VRAM, use Q4_K_M (file size 21.38 GB, estimated 22.3 GB memory). For 24 GB VRAM, use Q5_K_M (file size 24.75 GB, estimated 25.6 GB memory).

What GPU do you need for Granite 34b Code Base 8k?

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: no quant fits at an 8K context.
  • 12 GB VRAM: no quant fits at an 8K context.
  • 16 GB VRAM: the best quant that fits is IQ3_S (14.8 GB file, about 15.7 GB in use).
  • 24 GB VRAM: the best quant that fits is Q4_K_M (21.4 GB file, about 22.3 GB in use).
  • 32 GB VRAM: the best quant that fits is Q6_K (27.8 GB file, about 28.7 GB in use).
  • 48 GB VRAM: the best quant that fits is Q8_0 (36.0 GB file, about 36.9 GB in use).

Which Granite 34b Code Base 8k quant should you choose?

Q4_K_M is the usual starting point for Granite 34b Code Base 8k: it balances file size and quality well. Here it is a 21.4 GB file.

The Q8_0 file is 36.0 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 13.1 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

Architecturestarcoder
Parameters33.7B
Layers88
Trained context8,192 tokens
KV cache (fp16)44 KB per token

Source

Quantized bymradermacher
FamilyGranite
Licenseapache-2.0
Downloads (HF)1,680
Last updated2024-09-04

Frequently asked questions about Granite 34b Code Base 8k

How much VRAM does Granite 34b Code Base 8k need?

About 22.3 GB at Q4_K_M with an 8K context window. That is the 21.4 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 Granite 34b Code Base 8k on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 14.0 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Granite 34b Code Base 8k GGUF?

It depends on the quant: Q2_K is 13.1 GB, IQ3_XS is 14.3 GB, IQ3_S is 14.8 GB, Q3_K_S is 14.8 GB, IQ3_M is 15.9 GB, Q3_K_M is 17.6 GB and 8 more.

What context length does Granite 34b Code Base 8k support?

The GGUF header lists a trained context length of 8,192 tokens (8K). The KV cache grows with the context you use: about 0.4 GB at 8K and 0.4 GB at 8K.

Where can I download Granite 34b Code Base 8k GGUF?

The files are on Hugging Face in the mradermacher/granite-34b-code-base-8k-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 Granite 34b Code Base 8k use?

Hugging Face lists the license as apache-2.0. 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 Granite 34b Code Base 8k on 8 GB VRAM?

Yes, the Q2_K quant requires about 14 GB memory, which is close to 8 GB VRAM but may need CPU fallback or batching.

Does this model support vision tasks?

No, the model is a decoder‑only code model and has no vision capabilities.

Is this model suitable for new projects?

No, the model is deprecated and not recommended for new projects; newer Granite models should be used.

Other GGUFs of granite-34b-code-base-8k

Similar models

Browse all →

More from mradermacher

All by mradermacher →

More Granite models

All Granite →

Other medium models

All medium models →