aday777
/

deepseek_v4_tiny_fixture

Text Generation

Mixture of Experts

Model card Files Files and versions

deepseek_v4_tiny_fixture / generation_config.json

aday777's picture

Add tiny deterministic deepseek_v4 (DeepSeek-V4-Flash-Vision-Exp) loader/CI fixture: 150 tensors, 290,624 float32 params, MIT, donation bottom-only

42a3668 verified 4 days ago

History Blame Contribute Delete

94 Bytes

	{
	"bos_token_id": 0,
	"eos_token_id": 1,
	"no_repeat_ngram_size": 4,
	"seed": 20260903
	}