baidu
ERNIE-5.5-Flash-Base-BF16
text generationother618 downloadsHugging Face
No server in stock right now has the 640 GB this model needs for inference.
GPU memory it needs (estimate)
Rule of thumb from the model's size. Real needs depend on batch size, context length and framework.
Inference (full precision)
640 GB
Inference (8-bit)
320 GB
Inference (4-bit)
160 GB
LoRA fine-tuning
693 GB
QLoRA (8-bit base)
347 GB
Full fine-tuning
2559 GB