inference-optimization/test_tencentbac_fastmtp
Updated • 38
test-models
-
-
inference-optimization/test_qwen3_next_mtp
Updated • 41 -
inference-optimization/Qwen3-Next-80B-A3B-Instruct_mtp_speculator
Text Generation • 2B • Updated • 58 -
inference-optimization/Qwen3-Next-80B-A3B-Instruct-MTP-ultrachat-epoch3
2B • Updated • 18 -
inference-optimization/Qwen3-Next-80B-A3B-Instruct-MTP-ultrachat-epoch1
2B • Updated • 13 -
inference-optimization/Qwen3-Next-80B-A3B-Instruct-MTP-ultrachat-epoch2
2B • Updated • 10
inference-optimization/Qwen3-Next-80B-A3B-Instruct-GSM8K-MTP-finetuned
81B • Updated • 35Note Qwen3-Next-80B with GSM8K-finetuned MTP head (+22.5% acceptance rate). Generated using speculators.