GL

glm-4-9b-chat

GLMGGUF by sydneyfongUpdated 2024-07-10
Download on Hugging Face
Parameters9.4B
Context128K
Layers40
Architecturechatglm
Smallest quant4.6 GB
Quick answer

glm-4-9b-chat needs about 7.1 GB of memory at Q4_K_M (a 6.3 GB file) with an 8K context window, so it fits on a 8 GB GPU. This repository has 7 quants, from 4.6 GB (IQ3_S) to 10.0 GB (Q8_0). Longer context needs more: at 32K tokens the KV cache alone adds about 1.3 GB.

glm-4-9b-chat is a 9.4B-parameter GLM model in GGUF format, quantized by sydneyfong. Its trained context length is 128K tokens. The repository has 7 quant files, from 4.6 GB (IQ3_S) to 10.0 GB (Q8_0).

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ3_S4.6 GB 5.4 GB
Fits Get file →
Q4_K_S5.8 GB 6.6 GB
Fits Get file →
Q4_K_M6.3 GB 7.1 GB
Fits Get file →
Q5_K_S6.7 GB 7.5 GB
Tight Get file →
Q5_K_M7.1 GB 8.0 GB
Tight Get file →
Q6_K8.3 GB 9.1 GB
Too big Get file →
Q8_010.0 GB 10.8 GB
Too big Get file →

What GPU do you need for glm-4-9b-chat?

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: the best quant that fits is IQ3_S (4.6 GB file, about 5.4 GB in use).
  • 8 GB VRAM: the best quant that fits is Q5_K_M (7.1 GB file, about 8.0 GB in use).
  • 12 GB VRAM: the best quant that fits is Q8_0 (10.0 GB file, about 10.8 GB in use).

Which glm-4-9b-chat quant should you choose?

Q4_K_M is the usual starting point for glm-4-9b-chat: it balances file size and quality well. Here it is a 6.3 GB file.

The Q8_0 file is 10.0 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 IQ3_S at 4.6 GB, with the quality trade-offs that come with fewer bits. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturechatglm
Parameters9.4B
Layers40
Trained context131,072 tokens
KV cache (fp16)40 KB per token

Source

Quantized bysydneyfong
FamilyGLM
Licenseunknown
Downloads (HF)83
Last updated2024-07-10

Frequently asked questions about glm-4-9b-chat

How much VRAM does glm-4-9b-chat need?

About 7.1 GB at Q4_K_M with an 8K context window. That is the 6.3 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 glm-4-9b-chat on an 8 GB GPU?

Yes, with Q5_K_M: the 7.1 GB file needs about 8.0 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of glm-4-9b-chat GGUF?

It depends on the quant: IQ3_S is 4.6 GB, Q4_K_S is 5.8 GB, Q4_K_M is 6.3 GB, Q5_K_S is 6.7 GB, Q5_K_M is 7.1 GB, Q6_K is 8.3 GB and 1 more.

What context length does glm-4-9b-chat support?

The GGUF header lists a trained context length of 131,072 tokens (128K). The KV cache grows with the context you use: about 0.3 GB at 8K and 1.3 GB at 32K.

Where can I download glm-4-9b-chat GGUF?

The files are on Hugging Face in the sydneyfong/glm4-9b-chat-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 glm-4-9b-chat use?

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

Similar models

Browse all →

More from sydneyfong

All by sydneyfong →

More GLM models

All GLM →

Other small models

All small models →