clocksmith commited on
Commit
7c3d30e
·
verified ·
1 Parent(s): 52c928b

Publish gemma-3-1b-it-q4k-ehf16-af32 RDRR artifact

Browse files
models/gemma-3-1b-it-q4k-ehf16-af32/manifest.json CHANGED
@@ -136,29 +136,33 @@
136
  "maxBytes": 134217728,
137
  "outputDtype": "f32"
138
  }
139
- }
 
 
 
 
140
  },
141
  "execution": {
142
  "kernels": {
143
  "embed": {
144
  "kernel": "gather_f16.wgsl",
145
  "entry": "main",
146
- "digest": "sha256:a4829f4067091c98ad6ebbc9b0744cdd5bbcd4fbf6092b2f7cc7f1098695860f"
147
  },
148
  "rmsnorm": {
149
  "kernel": "rmsnorm.wgsl",
150
  "entry": "main",
151
- "digest": "sha256:f516b3e4bde2015f2a207c3ca5b8c9820c7809fa8f8d0786f90c568e0f1ac077"
152
  },
153
  "gemv": {
154
  "kernel": "matmul_gemv_subgroup.wgsl",
155
  "entry": "main_vec4",
156
- "digest": "sha256:3cee3bed453b40c5564a751d2a917649e10ad52f5268e77cbfecfcee34780457"
157
  },
158
  "rope": {
159
  "kernel": "rope.wgsl",
160
  "entry": "main",
161
- "digest": "sha256:238e989aad7da731046cd54d7bb3c80a6f26daf5d7a8e435607161de902e5882"
162
  },
163
  "attn_decode": {
164
  "kernel": "attention_decode_online_f16kv.wgsl",
@@ -194,7 +198,7 @@
194
  "attn_small": {
195
  "kernel": "attention_small_f16kv.wgsl",
196
  "entry": "main",
197
- "digest": "sha256:7e3acc24b8b45294d18052c759319fe8a202d4ebfc4b653d4b904b245ae7e5c9",
198
  "precision": {
199
  "kvDtype": "f16"
200
  }
@@ -210,7 +214,7 @@
210
  "lm_head_gemv": {
211
  "kernel": "matmul_gemv_subgroup.wgsl",
212
  "entry": "main_multicol",
213
- "digest": "sha256:96c38c15e6fed0d7efdc5cd094db5843a8e8ddfe01eee3bc7322fa555dacf3d0",
214
  "constants": {
215
  "MULTICOL_COLS_PER_WG": 64,
216
  "MULTICOL_THREADS_PER_COL": 4
@@ -219,7 +223,7 @@
219
  "sample": {
220
  "kernel": "sample.wgsl",
221
  "entry": "sample_single_pass",
222
- "digest": "sha256:4412357e84113ee2f1bc0dc8bf89e314c2ab482c89c14ca016ea9949d16a9d0c"
223
  }
224
  },
225
  "preLayer": [
@@ -392,6 +396,9 @@
392
  "dtypeTransition": "require_cast_step",
393
  "unresolvedKernel": "error"
394
  }
 
 
 
395
  }
396
  },
397
  "shards": [
@@ -4417,5 +4424,11 @@
4417
  "vocabSize": 262144,
4418
  "file": "tokenizer.json",
4419
  "addBosToken": true
 
 
 
 
 
 
4420
  }
4421
- }
 
136
  "maxBytes": 134217728,
137
  "outputDtype": "f32"
138
  }
139
+ },
140
+ "prefillChunkSubmitMode": "sync",
141
+ "useFlashPrefillAttention": false,
142
+ "useWideTileQ4KPrefill": false,
143
+ "retainQ4KMaterialization": false
144
  },
145
  "execution": {
146
  "kernels": {
147
  "embed": {
148
  "kernel": "gather_f16.wgsl",
149
  "entry": "main",
150
+ "digest": "sha256:e9f8056a8764360ca23e53289a5f133c2e3d6da122f6437f5eb0d78a46e8f049"
151
  },
152
  "rmsnorm": {
153
  "kernel": "rmsnorm.wgsl",
154
  "entry": "main",
155
+ "digest": "sha256:c723c0300cc8f93396d79d1a274c0a90ccf1eb8ce82c0d778b8cd4659615fe45"
156
  },
157
  "gemv": {
158
  "kernel": "matmul_gemv_subgroup.wgsl",
159
  "entry": "main_vec4",
160
+ "digest": "sha256:2ea9aea659367210fbf61d922c0d4ee17cdac68ffb506b3bc5516e498071e011"
161
  },
162
  "rope": {
163
  "kernel": "rope.wgsl",
164
  "entry": "main",
165
+ "digest": "sha256:e13c54f261d7ef00474b71e4fe7b3298fb5cc49161f50c7ea1854aa6672ebe71"
166
  },
167
  "attn_decode": {
168
  "kernel": "attention_decode_online_f16kv.wgsl",
 
198
  "attn_small": {
199
  "kernel": "attention_small_f16kv.wgsl",
200
  "entry": "main",
201
+ "digest": "sha256:369aa3b386b4ef89a916ba2717465caa47079f52bbe0f3e5d61bd3220e2ec71e",
202
  "precision": {
203
  "kvDtype": "f16"
204
  }
 
214
  "lm_head_gemv": {
215
  "kernel": "matmul_gemv_subgroup.wgsl",
216
  "entry": "main_multicol",
217
+ "digest": "sha256:e41f94574d5ac54dd2710036da5d5acc643a483b79b2d74b86825efdaaa7438f",
218
  "constants": {
219
  "MULTICOL_COLS_PER_WG": 64,
220
  "MULTICOL_THREADS_PER_COL": 4
 
223
  "sample": {
224
  "kernel": "sample.wgsl",
225
  "entry": "sample_single_pass",
226
+ "digest": "sha256:7ff3dd675b4737d920e0b3889e15c4fedde8410a353bbc6d4651db76ce69eb9b"
227
  }
228
  },
229
  "preLayer": [
 
396
  "dtypeTransition": "require_cast_step",
397
  "unresolvedKernel": "error"
398
  }
399
+ },
400
+ "largeWeights": {
401
+ "gpuResidentOverrides": null
402
  }
403
  },
404
  "shards": [
 
4424
  "vocabSize": 262144,
4425
  "file": "tokenizer.json",
4426
  "addBosToken": true
4427
+ },
4428
+ "artifactIdentity": {
4429
+ "sourceCheckpointId": "google/gemma-3-1b-it",
4430
+ "weightPackId": "gemma-3-1b-it-q4k-ehf16-af32-wp-catalog-v1",
4431
+ "manifestVariantId": "gemma-3-1b-it-q4k-ehf16-af32-mv-exec-v1",
4432
+ "artifactCompleteness": "complete"
4433
  }
4434
+ }