Configuration Parsing Warning:In config.json: "quantization_config.bits" must be an integer

mistralai / Mistral-Medium-3.5-128B-text

QUANTIZED BY: UnstableLlama
Information
Vision tower stripped to save space. Fits in 48gb VRAM with 16k context, kv quantized to 6,6
The KL is kind of high, so I verified it against another quant, and it measured in line.

2.50bpw exl3 quantization of Mistral-Medium-3.5-128B TEXT ONLY via exllamav3.
repo generated automatically with ezexl3.
Repo Data
REVISION GiB KL DIV PPL
2.25bpw 36.86 0.5566 7.1743
2.50bpw 39.61 0.4274 6.4994
bf16 119.44 0.0000 4.3142
CLI Download
hf download UnstableLlama/Mistral-Medium-3.5-128B-exl3-2.50bpw --local-dir ./Mistral-Medium-3.5-128B-exl3-2.50bpw
Downloads last month
9
Safetensors
Model size
21B params
Tensor type
BF16
F16
I16
Inference Providers NEW
This model isn't deployed by any Inference Provider. 馃檵 Ask for provider support