CR

CroissantLLMBase

CroissantllmbaseGGUF by mradermacherUpdated 2024-08-30
Download on Hugging Face
Parameters1.35B
Context2K
Layers24
Architecturellama
Smallest quant0.6 GB
Quick answer

CroissantLLMBase needs about 3.0 GB of memory at Q4_K_M (a 0.9 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 14 quants, from 0.6 GB (Q2_K) to 1.4 GB (Q8_0).

1.35B bilingual LLM with efficient GGUF quantizations

CroissantLLMBase is a 1.35B parameter bilingual model trained on 3T English and French tokens. It supports few‑shot prompting and runs on consumer‑grade hardware. The model is released under the MIT license.

Good at

  • Coding tasks
  • Bilingual text generation
  • Few‑shot prompting

Limits

  • No vision support
  • Context limited to 2048 tokens
  • Requires GGUF quantization

Best for

  • Local code generation
  • Bilingual content creation
  • Educational demos

Who it’s for. Developers and researchers needing a lightweight bilingual LLM for local deployment. The model is ideal for those who prefer open‑source solutions and have limited GPU memory.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q2_K0.6 GB 1.5 GB
Fits Get file →
IQ3_XS0.6 GB 1.5 GB
Fits Get file →
IQ3_S0.6 GB 1.5 GB
Fits Get file →
Q3_K_S0.6 GB 1.5 GB
Fits Get file →
IQ3_M0.7 GB 1.6 GB
Fits Get file →
Q3_K_M0.7 GB 1.6 GB
Fits Get file →
Q3_K_L0.7 GB 1.6 GB
Fits Get file →
IQ4_XS0.7 GB 1.7 GB
Fits Get file →
Q4_K_S0.8 GB 1.7 GB
Fits Get file →
Q4_K_M0.9 GB 1.8 GB
Fits Get file →
Q5_K_S1.0 GB 1.9 GB
Fits Get file →
Q5_K_M1.0 GB 1.9 GB
Fits Get file →
Q6_K1.2 GB 2.1 GB
Fits Get file →
Q8_01.4 GB 2.3 GB
Fits Get file →

For 8 GB VRAM, use Q4_K_M (0.87 GB file, ~3 GB memory at 8k context). For 4 GB VRAM, use Q3_K_S (0.64 GB file, ~2.8 GB memory). For 2 GB VRAM, use Q2_K (0.56 GB file, ~2.7 GB memory). Estimates are approximate.

What GPU do you need for CroissantLLMBase?

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 Q8_0 (1.4 GB file, about 3.5 GB in use).

Which CroissantLLMBase quant should you choose?

Q4_K_M is the usual starting point for CroissantLLMBase: it balances file size and quality well. Here it is a 0.9 GB file.

The Q8_0 file is 1.4 GB, about 1.6 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 0.6 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
Parameters1.35B
Layers24
Trained context2,048 tokens
KV cache (fp16)192 KB per token

Source

Quantized bymradermacher
Licensemit
Downloads (HF)1,614
Last updated2024-08-30

Frequently asked questions about CroissantLLMBase

How much VRAM does CroissantLLMBase need?

About 3.0 GB at Q4_K_M with an 8K context window. That is the 0.9 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 CroissantLLMBase on an 8 GB GPU?

Yes, with Q8_0: the 1.4 GB file needs about 3.5 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of CroissantLLMBase GGUF?

It depends on the quant: Q2_K is 0.6 GB, IQ3_XS is 0.6 GB, IQ3_S is 0.6 GB, Q3_K_S is 0.6 GB, IQ3_M is 0.7 GB, Q3_K_M is 0.7 GB and 8 more.

What context length does CroissantLLMBase support?

The GGUF header lists a trained context length of 2,048 tokens (2K). The KV cache grows with the context you use: about 1.6 GB at 8K and 0.4 GB at 2K.

Where can I download CroissantLLMBase GGUF?

The files are on Hugging Face in the mradermacher/CroissantLLMBase-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 CroissantLLMBase use?

Hugging Face lists the license as mit. 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 CroissantLLMBase on 4 GB VRAM?

Yes, using Q3_K_S or Q3_K_M quantizations will keep memory usage below 4 GB.

Does the model support French language tasks?

Yes, the model was trained on a 1:1 English‑to‑French token ratio and includes a French tokenizer.

Is there a chat version of this model?

The base model is not finetuned for chat; use the CroissantLLMChat-v0.1 for chat capabilities.

Similar models

Browse all →

More from mradermacher

All by mradermacher →

More Croissantllmbase models

All Croissantllmbase →

Other tiny models

All tiny models →