Initial K3 snapshot: 0.5B KDA/MLA/MoE train path

Standalone tree split from LLMRL/projects/kda. Includes Triton dt_bias
backward fix, train_k3 --preset 0.5b, SFT, Docker runtime, and tests.
This commit is contained in:
dela
2026-08-25 14:43:17 +08:00
commit 584f7e9e73
140 changed files with 21592 additions and 0 deletions
+10
View File
@@ -0,0 +1,10 @@
# Copyright (c) 2023-2026, Songlin Yang, Yu Zhang, Zhiyuan Li
#
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
# For a list of all contributors, visit:
# https://github.com/fla-org/flash-linear-attention/graphs/contributors
# Approximate value of 1/ln(2), used for log/exp base conversion
# Best FP32 approximation: 1.4426950216 (hex 0x3FB8AA3B)
RCP_LN2 = 1.4426950216