xhcao 9c6ee89fa7 [js/webgpu] fix two errors of attention operator (#21687) ... Fix two issues: (1) scale shall be fp32 instead of f16 (2) Softmax program does not handle the normalized dispatch group values, so if the sequence length is over 65535, the result is not correct for this program.		2024-08-13 09:42:34 -07:00
..
ops	[js/webgpu] fix two errors of attention operator (#21687)	2024-08-13 09:42:34 -07:00
attribute-with-cache-key.ts	[js] optimize eslint config (#18460)	2023-11-20 12:00:56 -08:00
gpu-data-manager.ts	more conservitive gpu-buffer cache algo (#20312)	2024-04-23 09:07:04 -07:00
op-resolve-rules.ts	[JS/WebGPU] Add Dequantizelinear operator (#21642)	2024-08-09 14:44:19 -07:00
program-manager.ts	[js/webgpu] add validation to workgroup size (#20110)	2024-04-02 19:29:20 -07:00
types.ts	[JS/WebGPU] Improve MatMulNBits perf (#19974)	2024-04-12 11:03:05 -07:00