Skip to content

Actions: hipudding/llama.cpp

Actions

Check Pre-Tokenizer Hashes

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
27 workflow runs
27 workflow runs

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
cann: upgrade flash attention from V2 to V5 for sinks and asymmetric … Check Pre-Tokenizer Hashes #27: Commit 09b2345 pushed by hipudding
1d 0h 0m 1s FA
FA
1d 0h 0m 1s
CANN: add new ops, optimize existing ops Check Pre-Tokenizer Hashes #26: Commit 01f5ebd pushed by hipudding
58s qwen35_op
qwen35_op
58s
Merge branch 'master' into qwen35_op Check Pre-Tokenizer Hashes #25: Commit a51b6d0 pushed by hipudding
55s qwen35_op
qwen35_op
55s
cann: implement backend memset_tensor interface Check Pre-Tokenizer Hashes #24: Commit c9dcafa pushed by hipudding
59s ocp
ocp
59s
CANN: fix multi-thread set_tensor race conditions Check Pre-Tokenizer Hashes #23: Commit aeee6b4 pushed by hipudding
1m 1s ocp
ocp
1m 1s
CANN: fix multi-thread set_tensor race conditions Check Pre-Tokenizer Hashes #22: Commit b9305e7 pushed by hipudding
59s ocp
ocp
59s
CANN: Support more functions in FA operator Check Pre-Tokenizer Hashes #21: Commit 1c219af pushed by hipudding
59s hua
hua
59s
CANN: implement quantized MUL_MAT_ID for MoE models Check Pre-Tokenizer Hashes #20: Commit 2d3679b pushed by hipudding
1m 1s mul_mat_id
mul_mat_id
1m 1s
Fix mul_mat_id_quant: use contiguous [M,D] layout for weight tensors Check Pre-Tokenizer Hashes #19: Commit e9d05be pushed by hipudding
57s mul_mat_id
mul_mat_id
57s
docs: update ops.md for CANN backend Check Pre-Tokenizer Hashes #18: Commit 0cd2858 pushed by hipudding
34s docs
docs
34s
Remove unused comments Check Pre-Tokenizer Hashes #17: Commit bfa67a8 pushed by hipudding
31s gla
gla
31s
CANN: Fix rename for get_env Check Pre-Tokenizer Hashes #16: Commit e25ac05 pushed by hipudding
32s post_fix_for_getenv
post_fix_for_getenv
32s
Resolve review comments Check Pre-Tokenizer Hashes #15: Commit 8dd5470 pushed by hipudding
30s vl
vl
30s
Check Pre-Tokenizer Hashes Check Pre-Tokenizer Hashes #14: Commit b2bc76c pushed by hipudding
33s vl
vl
33s
Check Pre-Tokenizer Hashes Check Pre-Tokenizer Hashes #13: Commit e173ee4 pushed by hipudding
32s unique_ptr
unique_ptr
32s
CANN: Update several operators to support FP16 data format Check Pre-Tokenizer Hashes #12: Commit bc59bb2 pushed by hipudding
51s fp16
fp16
51s
Support FP16 as intermediate results in graph computation Check Pre-Tokenizer Hashes #11: Commit f9a9eb1 pushed by hipudding
48s fp16_ext
fp16_ext
48s
Check Pre-Tokenizer Hashes Check Pre-Tokenizer Hashes #10: Commit 76c3ce5 pushed by hipudding
50s fp16_ext
fp16_ext
50s
Check Pre-Tokenizer Hashes Check Pre-Tokenizer Hashes #9: Commit 0b7c3cb pushed by hipudding
52s fp16
fp16
52s
fp16 Check Pre-Tokenizer Hashes #8: Commit 69492cc pushed by hipudding
1m 1s fp16
fp16
1m 1s
Check Pre-Tokenizer Hashes Check Pre-Tokenizer Hashes #7: Commit a6f67d0 pushed by hipudding
45s fp16
fp16
45s
CANN: Disable acl_graph for prefill stage Check Pre-Tokenizer Hashes #6: Commit 5c39f0c pushed by hipudding
53s prefill
prefill
53s
CANN: support ext_factor in rope Check Pre-Tokenizer Hashes #5: Commit cac9068 pushed by hipudding
29s rope_ext_factor
rope_ext_factor
29s
Check Pre-Tokenizer Hashes Check Pre-Tokenizer Hashes #4: Commit b2b1736 pushed by hipudding
29s rope_ext_factor
rope_ext_factor
29s
CANN: fix RoPE cache issue on multi-device Check Pre-Tokenizer Hashes #3: Commit c5e77de pushed by hipudding
32s rope_cache
rope_cache
32s