Skip to content

Commit 4955da0

Browse files
committed
Merge branch 'master' into xsn/mistral_large_moe
2 parents 2f8c2ef + 6016d0b commit 4955da0

39 files changed

+32977
-10634
lines changed

β€Ž.github/workflows/release.ymlβ€Ž

Lines changed: 4 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -67,7 +67,7 @@ jobs:
6767
run: |
6868
cp LICENSE ./build/bin/
6969
zip -y -r llama-${{ steps.tag.outputs.name }}-bin-macos-arm64.zip ./build/bin/*
70-
tar -czvf llama-${{ steps.tag.outputs.name }}-bin-macos-arm64.tar.gz -C ./build/bin .
70+
tar -czvf llama-${{ steps.tag.outputs.name }}-bin-macos-arm64.tar.gz -s ",./,llama-${{ steps.tag.outputs.name }}/," -C ./build/bin .
7171
7272
- name: Upload artifacts (zip)
7373
uses: actions/upload-artifact@v4
@@ -128,7 +128,7 @@ jobs:
128128
run: |
129129
cp LICENSE ./build/bin/
130130
zip -y -r llama-${{ steps.tag.outputs.name }}-bin-macos-x64.zip ./build/bin/*
131-
tar -czvf llama-${{ steps.tag.outputs.name }}-bin-macos-x64.tar.gz -C ./build/bin .
131+
tar -czvf llama-${{ steps.tag.outputs.name }}-bin-macos-x64.tar.gz -s ",./,llama-${{ steps.tag.outputs.name }}/," -C ./build/bin .
132132
133133
- name: Upload artifacts (zip)
134134
uses: actions/upload-artifact@v4
@@ -197,7 +197,7 @@ jobs:
197197
run: |
198198
cp LICENSE ./build/bin/
199199
zip -y -r llama-${{ steps.tag.outputs.name }}-bin-ubuntu-${{ matrix.build }}.zip ./build/bin/*
200-
tar -czvf llama-${{ steps.tag.outputs.name }}-bin-ubuntu-${{ matrix.build }}.tar.gz -C ./build/bin .
200+
tar -czvf llama-${{ steps.tag.outputs.name }}-bin-ubuntu-${{ matrix.build }}.tar.gz --transform "s,./,llama-${{ steps.tag.outputs.name }}/," -C ./build/bin .
201201
202202
- name: Upload artifacts (zip)
203203
uses: actions/upload-artifact@v4
@@ -257,7 +257,7 @@ jobs:
257257
run: |
258258
cp LICENSE ./build/bin/
259259
zip -y -r llama-${{ steps.tag.outputs.name }}-bin-ubuntu-vulkan-x64.zip ./build/bin/*
260-
tar -czvf llama-${{ steps.tag.outputs.name }}-bin-ubuntu-vulkan-x64.tar.gz -C ./build/bin .
260+
tar -czvf llama-${{ steps.tag.outputs.name }}-bin-ubuntu-vulkan-x64.tar.gz --transform "s,./,llama-${{ steps.tag.outputs.name }}/," -C ./build/bin .
261261
262262
- name: Upload artifacts (zip)
263263
uses: actions/upload-artifact@v4

β€Ž.github/workflows/winget.ymlβ€Ž

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -9,7 +9,7 @@ jobs:
99
update:
1010
name: Update Winget Package
1111
runs-on: ubuntu-latest
12-
if: ${{ github.repository.owner.login == 'ggml-org' }}
12+
if: github.repository_owner == 'ggml-org'
1313

1414
steps:
1515
- name: Install cargo binstall

β€ŽCODEOWNERSβ€Ž

Lines changed: 2 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -10,6 +10,7 @@
1010
/common/arg.* @ggerganov
1111
/common/base64.hpp.* @ggerganov
1212
/common/build-info.* @ggerganov
13+
/common/chat.* @pwilkin
1314
/common/chat-peg-parser.* @aldehir
1415
/common/common.* @ggerganov
1516
/common/console.* @ggerganov
@@ -84,6 +85,7 @@
8485
/src/llama-vocab.* @CISC
8586
/src/models/ @CISC
8687
/tests/ @ggerganov
88+
/tests/test-chat-.* @pwilkin
8789
/tools/batched-bench/ @ggerganov
8890
/tools/main/ @ggerganov
8991
/tools/mtmd/ @ngxson

β€Ždocs/ops.mdβ€Ž

Lines changed: 23 additions & 23 deletions
Original file line numberDiff line numberDiff line change
@@ -18,32 +18,32 @@ Legend:
1818
| ACC | ❌ | βœ… | βœ… | βœ… | βœ… | ❌ | βœ… | βœ… | ❌ |
1919
| ADD | ❌ | βœ… | βœ… | βœ… | 🟑 | 🟑 | βœ… | βœ… | ❌ |
2020
| ADD1 | ❌ | βœ… | βœ… | βœ… | ❌ | ❌ | βœ… | βœ… | ❌ |
21-
| ADD_ID | ❌ | ❌ | βœ… | βœ… | ❌ | ❌ | ❌ | βœ… | ❌ |
21+
| ADD_ID | ❌ | ❌ | βœ… | βœ… | βœ… | ❌ | ❌ | βœ… | ❌ |
2222
| ARANGE | ❌ | βœ… | βœ… | βœ… | βœ… | ❌ | βœ… | βœ… | ❌ |
2323
| ARGMAX | ❌ | βœ… | βœ… | βœ… | βœ… | ❌ | βœ… | βœ… | ❌ |
2424
| ARGSORT | ❌ | βœ… | βœ… | βœ… | βœ… | βœ… | βœ… | βœ… | ❌ |
2525
| CEIL | ❌ | ❌ | βœ… | 🟑 | ❌ | ❌ | 🟑 | 🟑 | ❌ |
2626
| CLAMP | ❌ | βœ… | βœ… | βœ… | 🟑 | 🟑 | 🟑 | 🟑 | ❌ |
2727
| CONCAT | ❌ | βœ… | βœ… | 🟑 | βœ… | 🟑 | βœ… | βœ… | ❌ |
2828
| CONT | ❌ | 🟑 | βœ… | βœ… | βœ… | 🟑 | 🟑 | βœ… | ❌ |
29-
| CONV_2D | ❌ | ❌ | βœ… | βœ… | ❌ | βœ… | ❌ | βœ… | ❌ |
29+
| CONV_2D | ❌ | ❌ | βœ… | βœ… | βœ… | βœ… | ❌ | βœ… | ❌ |
3030
| CONV_2D_DW | ❌ | ❌ | βœ… | βœ… | ❌ | ❌ | ❌ | βœ… | ❌ |
3131
| CONV_3D | ❌ | ❌ | βœ… | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ |
3232
| CONV_TRANSPOSE_1D | ❌ | βœ… | βœ… | βœ… | βœ… | ❌ | βœ… | βœ… | ❌ |
33-
| CONV_TRANSPOSE_2D | ❌ | ❌ | βœ… | βœ… | ❌ | ❌ | ❌ | βœ… | ❌ |
33+
| CONV_TRANSPOSE_2D | ❌ | ❌ | βœ… | βœ… | βœ… | ❌ | ❌ | βœ… | ❌ |
3434
| COS | ❌ | βœ… | βœ… | βœ… | 🟑 | ❌ | 🟑 | 🟑 | ❌ |
3535
| COUNT_EQUAL | ❌ | βœ… | βœ… | βœ… | ❌ | ❌ | βœ… | βœ… | ❌ |
3636
| CPY | ❌ | 🟑 | 🟑 | 🟑 | 🟑 | 🟑 | 🟑 | 🟑 | ❌ |
3737
| CROSS_ENTROPY_LOSS | ❌ | ❌ | βœ… | βœ… | ❌ | ❌ | ❌ | ❌ | ❌ |
3838
| CROSS_ENTROPY_LOSS_BACK | ❌ | ❌ | βœ… | βœ… | ❌ | ❌ | ❌ | ❌ | ❌ |
39-
| CUMSUM | ❌ | ❌ | βœ… | ❌ | ❌ | ❌ | ❌ | βœ… | ❌ |
40-
| DIAG_MASK_INF | ❌ | βœ… | βœ… | βœ… | 🟑 | 🟑 | βœ… | βœ… | ❌ |
39+
| CUMSUM | ❌ | ❌ | βœ… | ❌ | βœ… | ❌ | ❌ | βœ… | ❌ |
40+
| DIAG_MASK_INF | ❌ | βœ… | βœ… | βœ… | ❌ | 🟑 | βœ… | βœ… | ❌ |
4141
| DIV | ❌ | βœ… | βœ… | βœ… | 🟑 | 🟑 | βœ… | βœ… | ❌ |
4242
| DUP | ❌ | βœ… | βœ… | 🟑 | 🟑 | 🟑 | βœ… | βœ… | ❌ |
4343
| ELU | ❌ | βœ… | βœ… | 🟑 | 🟑 | ❌ | βœ… | ❌ | ❌ |
4444
| EXP | ❌ | βœ… | βœ… | 🟑 | 🟑 | ❌ | βœ… | 🟑 | ❌ |
45-
| EXPM1 | ❌ | ❌ | βœ… | 🟑 | ❌ | ❌ | ❌ | ❌ | ❌ |
46-
| FILL | ❌ | ❌ | βœ… | ❌ | ❌ | ❌ | ❌ | βœ… | ❌ |
45+
| EXPM1 | ❌ | ❌ | βœ… | 🟑 | 🟑 | ❌ | ❌ | ❌ | ❌ |
46+
| FILL | ❌ | ❌ | βœ… | ❌ | βœ… | ❌ | ❌ | βœ… | ❌ |
4747
| FLASH_ATTN_EXT | ❌ | 🟑 | βœ… | 🟑 | 🟑 | ❌ | ❌ | 🟑 | ❌ |
4848
| FLOOR | ❌ | ❌ | βœ… | 🟑 | ❌ | ❌ | 🟑 | 🟑 | ❌ |
4949
| GATED_LINEAR_ATTN | ❌ | ❌ | βœ… | βœ… | ❌ | ❌ | βœ… | ❌ | ❌ |
@@ -59,47 +59,47 @@ Legend:
5959
| GROUP_NORM_MUL_ADD | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ |
6060
| HARDSIGMOID | ❌ | βœ… | βœ… | 🟑 | 🟑 | ❌ | βœ… | 🟑 | ❌ |
6161
| HARDSWISH | ❌ | βœ… | βœ… | 🟑 | 🟑 | ❌ | βœ… | 🟑 | ❌ |
62-
| IM2COL | ❌ | βœ… | βœ… | βœ… | 🟑 | βœ… | βœ… | βœ… | ❌ |
62+
| IM2COL | ❌ | βœ… | βœ… | βœ… | βœ… | βœ… | βœ… | βœ… | ❌ |
6363
| IM2COL_3D | ❌ | ❌ | βœ… | βœ… | ❌ | ❌ | ❌ | βœ… | ❌ |
6464
| L2_NORM | ❌ | ❌ | βœ… | βœ… | βœ… | ❌ | βœ… | βœ… | ❌ |
65-
| LEAKY_RELU | ❌ | βœ… | βœ… | βœ… | βœ… | ❌ | βœ… | 🟑 | ❌ |
66-
| LOG | ❌ | βœ… | βœ… | βœ… | ❌ | ❌ | 🟑 | βœ… | ❌ |
65+
| LEAKY_RELU | ❌ | βœ… | βœ… | βœ… | 🟑 | ❌ | βœ… | 🟑 | ❌ |
66+
| LOG | ❌ | βœ… | βœ… | βœ… | 🟑 | ❌ | 🟑 | βœ… | ❌ |
6767
| MEAN | ❌ | βœ… | βœ… | βœ… | βœ… | ❌ | βœ… | βœ… | ❌ |
6868
| MUL | ❌ | βœ… | βœ… | βœ… | 🟑 | 🟑 | βœ… | βœ… | ❌ |
69-
| MUL_MAT | 🟑 | 🟑 | 🟑 | 🟑 | 🟑 | 🟑 | 🟑 | 🟑 | 🟑 |
69+
| MUL_MAT | 🟑 | 🟑 | 🟑 | 🟑 | βœ… | 🟑 | 🟑 | 🟑 | 🟑 |
7070
| MUL_MAT_ID | ❌ | 🟑 | βœ… | βœ… | βœ… | 🟑 | 🟑 | βœ… | ❌ |
7171
| NEG | ❌ | βœ… | βœ… | 🟑 | 🟑 | ❌ | βœ… | 🟑 | ❌ |
72-
| NORM | ❌ | βœ… | βœ… | βœ… | 🟑 | βœ… | βœ… | 🟑 | ❌ |
72+
| NORM | ❌ | βœ… | βœ… | βœ… | βœ… | βœ… | βœ… | 🟑 | ❌ |
7373
| NORM_MUL_ADD | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ |
74-
| OPT_STEP_ADAMW | ❌ | ❌ | βœ… | βœ… | ❌ | ❌ | ❌ | βœ… | ❌ |
75-
| OPT_STEP_SGD | ❌ | ❌ | βœ… | βœ… | ❌ | ❌ | ❌ | βœ… | ❌ |
74+
| OPT_STEP_ADAMW | ❌ | ❌ | βœ… | βœ… | βœ… | ❌ | ❌ | βœ… | ❌ |
75+
| OPT_STEP_SGD | ❌ | ❌ | βœ… | βœ… | βœ… | ❌ | ❌ | βœ… | ❌ |
7676
| OUT_PROD | 🟑 | ❌ | 🟑 | 🟑 | ❌ | ❌ | 🟑 | ❌ | ❌ |
77-
| PAD | ❌ | βœ… | βœ… | 🟑 | βœ… | βœ… | 🟑 | βœ… | ❌ |
77+
| PAD | ❌ | βœ… | βœ… | 🟑 | 🟑 | βœ… | 🟑 | βœ… | ❌ |
7878
| PAD_REFLECT_1D | ❌ | βœ… | βœ… | βœ… | βœ… | ❌ | βœ… | ❌ | ❌ |
7979
| POOL_2D | ❌ | 🟑 | βœ… | βœ… | βœ… | ❌ | βœ… | βœ… | ❌ |
8080
| REGLU | ❌ | βœ… | βœ… | βœ… | 🟑 | βœ… | βœ… | 🟑 | ❌ |
8181
| RELU | ❌ | βœ… | βœ… | 🟑 | 🟑 | 🟑 | βœ… | 🟑 | ❌ |
8282
| REPEAT | ❌ | βœ… | βœ… | 🟑 | βœ… | 🟑 | βœ… | 🟑 | ❌ |
8383
| REPEAT_BACK | ❌ | ❌ | βœ… | βœ… | ❌ | ❌ | βœ… | βœ… | ❌ |
84-
| RMS_NORM | ❌ | βœ… | βœ… | βœ… | 🟑 | βœ… | βœ… | βœ… | ❌ |
84+
| RMS_NORM | ❌ | βœ… | βœ… | βœ… | βœ… | βœ… | βœ… | βœ… | ❌ |
8585
| RMS_NORM_BACK | ❌ | ❌ | βœ… | βœ… | ❌ | ❌ | βœ… | βœ… | ❌ |
86-
| RMS_NORM_MUL_ADD | ❌ | βœ… | ❌ | ❌ | βœ… | βœ… | ❌ | ❌ | ❌ |
86+
| RMS_NORM_MUL_ADD | ❌ | βœ… | ❌ | ❌ | ❌ | βœ… | ❌ | ❌ | ❌ |
8787
| ROLL | ❌ | ❌ | βœ… | βœ… | ❌ | ❌ | βœ… | βœ… | ❌ |
8888
| ROPE | ❌ | 🟑 | βœ… | βœ… | βœ… | βœ… | βœ… | βœ… | ❌ |
8989
| ROPE_BACK | ❌ | ❌ | βœ… | βœ… | ❌ | ❌ | ❌ | βœ… | ❌ |
9090
| ROUND | ❌ | ❌ | βœ… | 🟑 | ❌ | ❌ | 🟑 | 🟑 | ❌ |
9191
| RWKV_WKV6 | ❌ | ❌ | βœ… | βœ… | βœ… | ❌ | βœ… | βœ… | ❌ |
9292
| RWKV_WKV7 | ❌ | ❌ | βœ… | βœ… | βœ… | ❌ | βœ… | βœ… | ❌ |
9393
| SCALE | ❌ | 🟑 | βœ… | βœ… | βœ… | βœ… | βœ… | βœ… | ❌ |
94-
| SET | ❌ | ❌ | βœ… | βœ… | βœ… | ❌ | 🟑 | ❌ | ❌ |
94+
| SET | ❌ | ❌ | βœ… | βœ… | ❌ | ❌ | 🟑 | ❌ | ❌ |
9595
| SET_ROWS | ❌ | ❌ | 🟑 | 🟑 | 🟑 | 🟑 | 🟑 | 🟑 | ❌ |
9696
| SGN | ❌ | βœ… | βœ… | 🟑 | 🟑 | ❌ | βœ… | ❌ | ❌ |
9797
| SIGMOID | ❌ | βœ… | βœ… | 🟑 | 🟑 | 🟑 | βœ… | 🟑 | ❌ |
9898
| SILU | ❌ | βœ… | βœ… | 🟑 | 🟑 | 🟑 | βœ… | 🟑 | ❌ |
9999
| SILU_BACK | ❌ | ❌ | βœ… | βœ… | ❌ | ❌ | ❌ | βœ… | ❌ |
100100
| SIN | ❌ | βœ… | βœ… | βœ… | 🟑 | ❌ | 🟑 | 🟑 | ❌ |
101101
| SOFTCAP | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ |
102-
| SOFTPLUS | ❌ | ❌ | βœ… | 🟑 | ❌ | ❌ | ❌ | 🟑 | ❌ |
102+
| SOFTPLUS | ❌ | ❌ | βœ… | 🟑 | 🟑 | ❌ | ❌ | 🟑 | ❌ |
103103
| SOFT_MAX | ❌ | 🟑 | βœ… | βœ… | βœ… | βœ… | βœ… | βœ… | ❌ |
104104
| SOFT_MAX_BACK | ❌ | ❌ | 🟑 | 🟑 | ❌ | ❌ | 🟑 | βœ… | ❌ |
105105
| SOLVE_TRI | ❌ | ❌ | βœ… | ❌ | ❌ | ❌ | ❌ | 🟑 | ❌ |
@@ -109,14 +109,14 @@ Legend:
109109
| SSM_SCAN | ❌ | ❌ | βœ… | βœ… | βœ… | ❌ | ❌ | 🟑 | ❌ |
110110
| STEP | ❌ | βœ… | βœ… | 🟑 | 🟑 | ❌ | βœ… | 🟑 | ❌ |
111111
| SUB | ❌ | βœ… | βœ… | βœ… | 🟑 | 🟑 | βœ… | βœ… | ❌ |
112-
| SUM | ❌ | βœ… | βœ… | 🟑 | ❌ | ❌ | 🟑 | 🟑 | ❌ |
112+
| SUM | ❌ | βœ… | βœ… | 🟑 | 🟑 | ❌ | 🟑 | 🟑 | ❌ |
113113
| SUM_ROWS | ❌ | βœ… | βœ… | 🟑 | βœ… | βœ… | 🟑 | βœ… | ❌ |
114114
| SWIGLU | ❌ | βœ… | βœ… | βœ… | 🟑 | βœ… | βœ… | 🟑 | ❌ |
115-
| SWIGLU_OAI | ❌ | ❌ | βœ… | βœ… | ❌ | ❌ | ❌ | 🟑 | ❌ |
115+
| SWIGLU_OAI | ❌ | ❌ | βœ… | βœ… | βœ… | ❌ | ❌ | 🟑 | ❌ |
116116
| TANH | ❌ | βœ… | βœ… | 🟑 | 🟑 | βœ… | βœ… | 🟑 | ❌ |
117117
| TIMESTEP_EMBEDDING | ❌ | βœ… | βœ… | βœ… | βœ… | βœ… | βœ… | βœ… | ❌ |
118-
| TOP_K | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | 🟑 | ❌ |
119-
| TRI | ❌ | ❌ | βœ… | ❌ | ❌ | ❌ | ❌ | βœ… | ❌ |
118+
| TOP_K | ❌ | ❌ | ❌ | ❌ | βœ… | ❌ | ❌ | 🟑 | ❌ |
119+
| TRI | ❌ | ❌ | βœ… | ❌ | βœ… | ❌ | ❌ | βœ… | ❌ |
120120
| TRUNC | ❌ | ❌ | βœ… | 🟑 | ❌ | ❌ | 🟑 | 🟑 | ❌ |
121121
| UPSCALE | ❌ | 🟑 | βœ… | βœ… | 🟑 | βœ… | 🟑 | 🟑 | ❌ |
122122
| XIELU | ❌ | ❌ | βœ… | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ |

0 commit comments

Comments
Β (0)