Upload folder using huggingface_hub
Browse files- layerwise_stats.csv +0 -18
- layerwise_std_granite-3.1-8b-base.pdf +0 -0
- layerwise_std_granite-3.3-8b-base.pdf +0 -0
- module_type_stats.csv +0 -18
- module_type_std_granite-3.1-8b-base.pdf +0 -0
- module_type_std_granite-3.3-8b-base.pdf +0 -0
- parameter_stats.csv +0 -0
- sublayer_stats.csv +0 -18
- sublayer_std_granite-3.1-8b-base.pdf +0 -0
- sublayer_std_granite-3.3-8b-base.pdf +0 -0
- summary.md +0 -0
layerwise_stats.csv
CHANGED
|
@@ -1,12 +1,4 @@
|
|
| 1 |
model,block_label,block_sort,count,mean,std,stat_mode,base_model
|
| 2 |
-
xplain1.3-6base-16k-step2199,Embedding,-1000000,201326592,2.3755903081716193e-07,0.0014961385525487114,signed,granite-3.1-8b-base
|
| 3 |
-
xplain1.3-6base-32k-step1647,Embedding,-1000000,201326592,1.1241835788193082e-07,0.001761873225711202,signed,granite-3.1-8b-base
|
| 4 |
-
xplain1.3-6base-64k-step1647,Embedding,-1000000,201326592,-3.4200752452306915e-07,0.002014212355859453,signed,granite-3.1-8b-base
|
| 5 |
-
xplain1.3-6base-8k-step2199,Embedding,-1000000,201326592,9.71491923943783e-07,0.0010270567109189673,signed,granite-3.1-8b-base
|
| 6 |
-
xplain1.4-6base-16k-step2319,Embedding,-1000000,201326592,2.4094494316765727e-06,0.001679461976228573,signed,granite-3.1-8b-base
|
| 7 |
-
xplain1.4-6base-32k-step1743,Embedding,-1000000,201326592,4.533886794888531e-07,0.002007301041444322,signed,granite-3.1-8b-base
|
| 8 |
-
xplain1.4-6base-64k-step1743,Embedding,-1000000,201326592,3.0080029015759163e-07,0.0022474787383138226,signed,granite-3.1-8b-base
|
| 9 |
-
xplain1.4-6base-8k-step2319,Embedding,-1000000,201326592,1.593794271078271e-06,0.0011852443582814004,signed,granite-3.1-8b-base
|
| 10 |
xplain1.3-6base-16k-step2199,Layer 00,0,199237632,-8.987953314599564e-08,0.0010252865255644995,signed,granite-3.1-8b-base
|
| 11 |
xplain1.3-6base-32k-step1647,Layer 00,0,199237632,4.7587162088732213e-08,0.0011208736232560896,signed,granite-3.1-8b-base
|
| 12 |
xplain1.3-6base-64k-step1647,Layer 00,0,199237632,1.0770314295216843e-07,0.0013730658491330555,signed,granite-3.1-8b-base
|
|
@@ -327,15 +319,6 @@ xplain1.4-6base-16k-step2319,Layer 39,39,199237632,-4.537560513974781e-07,0.0015
|
|
| 327 |
xplain1.4-6base-32k-step1743,Layer 39,39,199237632,-3.919972705337811e-07,0.0017778195017764222,signed,granite-3.1-8b-base
|
| 328 |
xplain1.4-6base-64k-step1743,Layer 39,39,199237632,-4.3374580167836904e-07,0.0019091703034062617,signed,granite-3.1-8b-base
|
| 329 |
xplain1.4-6base-8k-step2319,Layer 39,39,199237632,-3.4125923325103393e-07,0.0011943516630719838,signed,granite-3.1-8b-base
|
| 330 |
-
xplain1.3-6base-16k-step2199,Final Norm,900000,4096,-0.00042849575402215123,0.0038598968849905264,signed,granite-3.1-8b-base
|
| 331 |
-
xplain1.3-6base-32k-step1647,Final Norm,900000,4096,0.00010551830200711265,0.005347267957942419,signed,granite-3.1-8b-base
|
| 332 |
-
xplain1.3-6base-64k-step1647,Final Norm,900000,4096,0.0006408124463632703,0.006868563957013567,signed,granite-3.1-8b-base
|
| 333 |
-
xplain1.3-6base-8k-step2199,Final Norm,900000,4096,0.0015380806289613247,0.0023878312117495497,signed,granite-3.1-8b-base
|
| 334 |
-
xplain1.4-6base-16k-step2319,Final Norm,900000,4096,0.0071195634081959724,0.005594272826879382,signed,granite-3.1-8b-base
|
| 335 |
-
xplain1.4-6base-32k-step1743,Final Norm,900000,4096,0.005457671359181404,0.006660228220064173,signed,granite-3.1-8b-base
|
| 336 |
-
xplain1.4-6base-64k-step1743,Final Norm,900000,4096,0.004167107865214348,0.007889068659360296,signed,granite-3.1-8b-base
|
| 337 |
-
xplain1.4-6base-8k-step2319,Final Norm,900000,4096,0.0034991619177162647,0.003268739574482555,signed,granite-3.1-8b-base
|
| 338 |
-
xplain1.4-7base-32k-step1743,Embedding,-1000000,201326592,1.1008651199517772e-06,0.00171573377963928,signed,granite-3.3-8b-base
|
| 339 |
xplain1.4-7base-32k-step1743,Layer 00,0,199237632,-8.117254924692805e-08,0.0009165955813002411,signed,granite-3.3-8b-base
|
| 340 |
xplain1.4-7base-32k-step1743,Layer 01,1,199237632,-8.992852155742073e-08,0.0009609869916258348,signed,granite-3.3-8b-base
|
| 341 |
xplain1.4-7base-32k-step1743,Layer 02,2,199237632,6.131856044787931e-08,0.0010101430954313595,signed,granite-3.3-8b-base
|
|
@@ -376,4 +359,3 @@ xplain1.4-7base-32k-step1743,Layer 36,36,199237632,4.1582346968326754e-08,0.0015
|
|
| 376 |
xplain1.4-7base-32k-step1743,Layer 37,37,199237632,-2.485622219375417e-07,0.0015784436605922924,signed,granite-3.3-8b-base
|
| 377 |
xplain1.4-7base-32k-step1743,Layer 38,38,199237632,-4.2233954251085863e-07,0.0015787066318938686,signed,granite-3.3-8b-base
|
| 378 |
xplain1.4-7base-32k-step1743,Layer 39,39,199237632,-3.779794359035348e-07,0.0014695925431838642,signed,granite-3.3-8b-base
|
| 379 |
-
xplain1.4-7base-32k-step1743,Final Norm,900000,4096,0.012469591572880745,0.006104727263580456,signed,granite-3.3-8b-base
|
|
|
|
| 1 |
model,block_label,block_sort,count,mean,std,stat_mode,base_model
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
xplain1.3-6base-16k-step2199,Layer 00,0,199237632,-8.987953314599564e-08,0.0010252865255644995,signed,granite-3.1-8b-base
|
| 3 |
xplain1.3-6base-32k-step1647,Layer 00,0,199237632,4.7587162088732213e-08,0.0011208736232560896,signed,granite-3.1-8b-base
|
| 4 |
xplain1.3-6base-64k-step1647,Layer 00,0,199237632,1.0770314295216843e-07,0.0013730658491330555,signed,granite-3.1-8b-base
|
|
|
|
| 319 |
xplain1.4-6base-32k-step1743,Layer 39,39,199237632,-3.919972705337811e-07,0.0017778195017764222,signed,granite-3.1-8b-base
|
| 320 |
xplain1.4-6base-64k-step1743,Layer 39,39,199237632,-4.3374580167836904e-07,0.0019091703034062617,signed,granite-3.1-8b-base
|
| 321 |
xplain1.4-6base-8k-step2319,Layer 39,39,199237632,-3.4125923325103393e-07,0.0011943516630719838,signed,granite-3.1-8b-base
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 322 |
xplain1.4-7base-32k-step1743,Layer 00,0,199237632,-8.117254924692805e-08,0.0009165955813002411,signed,granite-3.3-8b-base
|
| 323 |
xplain1.4-7base-32k-step1743,Layer 01,1,199237632,-8.992852155742073e-08,0.0009609869916258348,signed,granite-3.3-8b-base
|
| 324 |
xplain1.4-7base-32k-step1743,Layer 02,2,199237632,6.131856044787931e-08,0.0010101430954313595,signed,granite-3.3-8b-base
|
|
|
|
| 359 |
xplain1.4-7base-32k-step1743,Layer 37,37,199237632,-2.485622219375417e-07,0.0015784436605922924,signed,granite-3.3-8b-base
|
| 360 |
xplain1.4-7base-32k-step1743,Layer 38,38,199237632,-4.2233954251085863e-07,0.0015787066318938686,signed,granite-3.3-8b-base
|
| 361 |
xplain1.4-7base-32k-step1743,Layer 39,39,199237632,-3.779794359035348e-07,0.0014695925431838642,signed,granite-3.3-8b-base
|
|
|
layerwise_std_granite-3.1-8b-base.pdf
CHANGED
|
Binary files a/layerwise_std_granite-3.1-8b-base.pdf and b/layerwise_std_granite-3.1-8b-base.pdf differ
|
|
|
layerwise_std_granite-3.3-8b-base.pdf
CHANGED
|
Binary files a/layerwise_std_granite-3.3-8b-base.pdf and b/layerwise_std_granite-3.3-8b-base.pdf differ
|
|
|
module_type_stats.csv
CHANGED
|
@@ -1,12 +1,4 @@
|
|
| 1 |
model,module_type,module_sort,count,mean,std,stat_mode,base_model
|
| 2 |
-
xplain1.3-6base-16k-step2199,embed_tokens,0,201326592,2.3755903081716193e-07,0.0014961385525487114,signed,granite-3.1-8b-base
|
| 3 |
-
xplain1.3-6base-32k-step1647,embed_tokens,0,201326592,1.1241835788193082e-07,0.001761873225711202,signed,granite-3.1-8b-base
|
| 4 |
-
xplain1.3-6base-64k-step1647,embed_tokens,0,201326592,-3.4200752452306915e-07,0.002014212355859453,signed,granite-3.1-8b-base
|
| 5 |
-
xplain1.3-6base-8k-step2199,embed_tokens,0,201326592,9.71491923943783e-07,0.0010270567109189673,signed,granite-3.1-8b-base
|
| 6 |
-
xplain1.4-6base-16k-step2319,embed_tokens,0,201326592,2.4094494316765727e-06,0.001679461976228573,signed,granite-3.1-8b-base
|
| 7 |
-
xplain1.4-6base-32k-step1743,embed_tokens,0,201326592,4.533886794888531e-07,0.002007301041444322,signed,granite-3.1-8b-base
|
| 8 |
-
xplain1.4-6base-64k-step1743,embed_tokens,0,201326592,3.0080029015759163e-07,0.0022474787383138226,signed,granite-3.1-8b-base
|
| 9 |
-
xplain1.4-6base-8k-step2319,embed_tokens,0,201326592,1.593794271078271e-06,0.0011852443582814004,signed,granite-3.1-8b-base
|
| 10 |
xplain1.3-6base-16k-step2199,self_attn.q_proj,2,671088640,7.45930424095321e-08,0.0011643161510070817,signed,granite-3.1-8b-base
|
| 11 |
xplain1.3-6base-32k-step1647,self_attn.q_proj,2,671088640,6.552511182533759e-08,0.0013388981678615203,signed,granite-3.1-8b-base
|
| 12 |
xplain1.3-6base-64k-step1647,self_attn.q_proj,2,671088640,8.005631055851846e-08,0.001510442626541124,signed,granite-3.1-8b-base
|
|
@@ -79,15 +71,6 @@ xplain1.4-6base-16k-step2319,post_attention_layernorm,10,163840,-0.0005267394903
|
|
| 79 |
xplain1.4-6base-32k-step1743,post_attention_layernorm,10,163840,-0.0006590992120436567,0.0016037598977734619,signed,granite-3.1-8b-base
|
| 80 |
xplain1.4-6base-64k-step1743,post_attention_layernorm,10,163840,-0.0008427060225130844,0.0017844306164154927,signed,granite-3.1-8b-base
|
| 81 |
xplain1.4-6base-8k-step2319,post_attention_layernorm,10,163840,-0.000276007588399807,0.0009058837622081566,signed,granite-3.1-8b-base
|
| 82 |
-
xplain1.3-6base-16k-step2199,final_norm,16,4096,-0.00042849575402215123,0.0038598968849905264,signed,granite-3.1-8b-base
|
| 83 |
-
xplain1.3-6base-32k-step1647,final_norm,16,4096,0.00010551830200711265,0.005347267957942419,signed,granite-3.1-8b-base
|
| 84 |
-
xplain1.3-6base-64k-step1647,final_norm,16,4096,0.0006408124463632703,0.006868563957013567,signed,granite-3.1-8b-base
|
| 85 |
-
xplain1.3-6base-8k-step2199,final_norm,16,4096,0.0015380806289613247,0.0023878312117495497,signed,granite-3.1-8b-base
|
| 86 |
-
xplain1.4-6base-16k-step2319,final_norm,16,4096,0.0071195634081959724,0.005594272826879382,signed,granite-3.1-8b-base
|
| 87 |
-
xplain1.4-6base-32k-step1743,final_norm,16,4096,0.005457671359181404,0.006660228220064173,signed,granite-3.1-8b-base
|
| 88 |
-
xplain1.4-6base-64k-step1743,final_norm,16,4096,0.004167107865214348,0.007889068659360296,signed,granite-3.1-8b-base
|
| 89 |
-
xplain1.4-6base-8k-step2319,final_norm,16,4096,0.0034991619177162647,0.003268739574482555,signed,granite-3.1-8b-base
|
| 90 |
-
xplain1.4-7base-32k-step1743,embed_tokens,0,201326592,1.1008651199517772e-06,0.00171573377963928,signed,granite-3.3-8b-base
|
| 91 |
xplain1.4-7base-32k-step1743,self_attn.q_proj,2,671088640,-5.0565586651174423e-08,0.0012394320928580485,signed,granite-3.3-8b-base
|
| 92 |
xplain1.4-7base-32k-step1743,self_attn.k_proj,3,167772160,1.3994926728599858e-08,0.0011508524608267654,signed,granite-3.3-8b-base
|
| 93 |
xplain1.4-7base-32k-step1743,self_attn.v_proj,4,167772160,-1.3449267299847634e-08,0.0009255093768879193,signed,granite-3.3-8b-base
|
|
@@ -97,4 +80,3 @@ xplain1.4-7base-32k-step1743,mlp.up_proj,7,2097152000,-1.0294117885223387e-08,0.
|
|
| 97 |
xplain1.4-7base-32k-step1743,mlp.down_proj,8,2097152000,4.27254578312386e-08,0.0013068294218553456,signed,granite-3.3-8b-base
|
| 98 |
xplain1.4-7base-32k-step1743,input_layernorm,9,163840,-0.0033162338767851905,0.0027314397751208356,signed,granite-3.3-8b-base
|
| 99 |
xplain1.4-7base-32k-step1743,post_attention_layernorm,10,163840,-0.0007859055350309063,0.00145576000964007,signed,granite-3.3-8b-base
|
| 100 |
-
xplain1.4-7base-32k-step1743,final_norm,16,4096,0.012469591572880745,0.006104727263580456,signed,granite-3.3-8b-base
|
|
|
|
| 1 |
model,module_type,module_sort,count,mean,std,stat_mode,base_model
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
xplain1.3-6base-16k-step2199,self_attn.q_proj,2,671088640,7.45930424095321e-08,0.0011643161510070817,signed,granite-3.1-8b-base
|
| 3 |
xplain1.3-6base-32k-step1647,self_attn.q_proj,2,671088640,6.552511182533759e-08,0.0013388981678615203,signed,granite-3.1-8b-base
|
| 4 |
xplain1.3-6base-64k-step1647,self_attn.q_proj,2,671088640,8.005631055851846e-08,0.001510442626541124,signed,granite-3.1-8b-base
|
|
|
|
| 71 |
xplain1.4-6base-32k-step1743,post_attention_layernorm,10,163840,-0.0006590992120436567,0.0016037598977734619,signed,granite-3.1-8b-base
|
| 72 |
xplain1.4-6base-64k-step1743,post_attention_layernorm,10,163840,-0.0008427060225130844,0.0017844306164154927,signed,granite-3.1-8b-base
|
| 73 |
xplain1.4-6base-8k-step2319,post_attention_layernorm,10,163840,-0.000276007588399807,0.0009058837622081566,signed,granite-3.1-8b-base
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 74 |
xplain1.4-7base-32k-step1743,self_attn.q_proj,2,671088640,-5.0565586651174423e-08,0.0012394320928580485,signed,granite-3.3-8b-base
|
| 75 |
xplain1.4-7base-32k-step1743,self_attn.k_proj,3,167772160,1.3994926728599858e-08,0.0011508524608267654,signed,granite-3.3-8b-base
|
| 76 |
xplain1.4-7base-32k-step1743,self_attn.v_proj,4,167772160,-1.3449267299847634e-08,0.0009255093768879193,signed,granite-3.3-8b-base
|
|
|
|
| 80 |
xplain1.4-7base-32k-step1743,mlp.down_proj,8,2097152000,4.27254578312386e-08,0.0013068294218553456,signed,granite-3.3-8b-base
|
| 81 |
xplain1.4-7base-32k-step1743,input_layernorm,9,163840,-0.0033162338767851905,0.0027314397751208356,signed,granite-3.3-8b-base
|
| 82 |
xplain1.4-7base-32k-step1743,post_attention_layernorm,10,163840,-0.0007859055350309063,0.00145576000964007,signed,granite-3.3-8b-base
|
|
|
module_type_std_granite-3.1-8b-base.pdf
CHANGED
|
Binary files a/module_type_std_granite-3.1-8b-base.pdf and b/module_type_std_granite-3.1-8b-base.pdf differ
|
|
|
module_type_std_granite-3.3-8b-base.pdf
CHANGED
|
Binary files a/module_type_std_granite-3.3-8b-base.pdf and b/module_type_std_granite-3.3-8b-base.pdf differ
|
|
|
parameter_stats.csv
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
sublayer_stats.csv
CHANGED
|
@@ -1,12 +1,4 @@
|
|
| 1 |
model,sublayer,sublayer_sort,block_label,block_sort,module_type,module_sort,count,mean,std,stat_mode,base_model
|
| 2 |
-
xplain1.3-6base-16k-step2199,Embedding | model.embed_tokens.weight,-1000000,Embedding,-1000000,embed_tokens,0,201326592,2.3755903081716193e-07,0.0014961385525487114,signed,granite-3.1-8b-base
|
| 3 |
-
xplain1.3-6base-32k-step1647,Embedding | model.embed_tokens.weight,-1000000,Embedding,-1000000,embed_tokens,0,201326592,1.1241835788193082e-07,0.001761873225711202,signed,granite-3.1-8b-base
|
| 4 |
-
xplain1.3-6base-64k-step1647,Embedding | model.embed_tokens.weight,-1000000,Embedding,-1000000,embed_tokens,0,201326592,-3.4200752452306915e-07,0.002014212355859453,signed,granite-3.1-8b-base
|
| 5 |
-
xplain1.3-6base-8k-step2199,Embedding | model.embed_tokens.weight,-1000000,Embedding,-1000000,embed_tokens,0,201326592,9.71491923943783e-07,0.0010270567109189673,signed,granite-3.1-8b-base
|
| 6 |
-
xplain1.4-6base-16k-step2319,Embedding | model.embed_tokens.weight,-1000000,Embedding,-1000000,embed_tokens,0,201326592,2.4094494316765727e-06,0.001679461976228573,signed,granite-3.1-8b-base
|
| 7 |
-
xplain1.4-6base-32k-step1743,Embedding | model.embed_tokens.weight,-1000000,Embedding,-1000000,embed_tokens,0,201326592,4.533886794888531e-07,0.002007301041444322,signed,granite-3.1-8b-base
|
| 8 |
-
xplain1.4-6base-64k-step1743,Embedding | model.embed_tokens.weight,-1000000,Embedding,-1000000,embed_tokens,0,201326592,3.0080029015759163e-07,0.0022474787383138226,signed,granite-3.1-8b-base
|
| 9 |
-
xplain1.4-6base-8k-step2319,Embedding | model.embed_tokens.weight,-1000000,Embedding,-1000000,embed_tokens,0,201326592,1.593794271078271e-06,0.0011852443582814004,signed,granite-3.1-8b-base
|
| 10 |
xplain1.3-6base-16k-step2199,Layer 00 | self_attn.q_proj,2,Layer 00,0,self_attn.q_proj,2,16777216,-2.5386498236912303e-08,0.001145262995459996,signed,granite-3.1-8b-base
|
| 11 |
xplain1.3-6base-32k-step1647,Layer 00 | self_attn.q_proj,2,Layer 00,0,self_attn.q_proj,2,16777216,2.6751570203487063e-07,0.001236013762140753,signed,granite-3.1-8b-base
|
| 12 |
xplain1.3-6base-64k-step1647,Layer 00 | self_attn.q_proj,2,Layer 00,0,self_attn.q_proj,2,16777216,3.582210297281563e-07,0.0014320992904867204,signed,granite-3.1-8b-base
|
|
@@ -2887,15 +2879,6 @@ xplain1.4-6base-16k-step2319,Layer 39 | post_attention_layernorm,39010,Layer 39,
|
|
| 2887 |
xplain1.4-6base-32k-step1743,Layer 39 | post_attention_layernorm,39010,Layer 39,39,post_attention_layernorm,10,4096,0.0011097422102466226,0.0020724020711700023,signed,granite-3.1-8b-base
|
| 2888 |
xplain1.4-6base-64k-step1743,Layer 39 | post_attention_layernorm,39010,Layer 39,39,post_attention_layernorm,10,4096,0.00044853618601337075,0.002316154528465667,signed,granite-3.1-8b-base
|
| 2889 |
xplain1.4-6base-8k-step2319,Layer 39 | post_attention_layernorm,39010,Layer 39,39,post_attention_layernorm,10,4096,0.0006851548096165061,0.0010263779705000592,signed,granite-3.1-8b-base
|
| 2890 |
-
xplain1.3-6base-16k-step2199,Final Norm | model.norm.weight,900000,Final Norm,900000,final_norm,16,4096,-0.00042849575402215123,0.0038598968849905264,signed,granite-3.1-8b-base
|
| 2891 |
-
xplain1.3-6base-32k-step1647,Final Norm | model.norm.weight,900000,Final Norm,900000,final_norm,16,4096,0.00010551830200711265,0.005347267957942419,signed,granite-3.1-8b-base
|
| 2892 |
-
xplain1.3-6base-64k-step1647,Final Norm | model.norm.weight,900000,Final Norm,900000,final_norm,16,4096,0.0006408124463632703,0.006868563957013567,signed,granite-3.1-8b-base
|
| 2893 |
-
xplain1.3-6base-8k-step2199,Final Norm | model.norm.weight,900000,Final Norm,900000,final_norm,16,4096,0.0015380806289613247,0.0023878312117495497,signed,granite-3.1-8b-base
|
| 2894 |
-
xplain1.4-6base-16k-step2319,Final Norm | model.norm.weight,900000,Final Norm,900000,final_norm,16,4096,0.0071195634081959724,0.005594272826879382,signed,granite-3.1-8b-base
|
| 2895 |
-
xplain1.4-6base-32k-step1743,Final Norm | model.norm.weight,900000,Final Norm,900000,final_norm,16,4096,0.005457671359181404,0.006660228220064173,signed,granite-3.1-8b-base
|
| 2896 |
-
xplain1.4-6base-64k-step1743,Final Norm | model.norm.weight,900000,Final Norm,900000,final_norm,16,4096,0.004167107865214348,0.007889068659360296,signed,granite-3.1-8b-base
|
| 2897 |
-
xplain1.4-6base-8k-step2319,Final Norm | model.norm.weight,900000,Final Norm,900000,final_norm,16,4096,0.0034991619177162647,0.003268739574482555,signed,granite-3.1-8b-base
|
| 2898 |
-
xplain1.4-7base-32k-step1743,Embedding | model.embed_tokens.weight,-1000000,Embedding,-1000000,embed_tokens,0,201326592,1.1008651199517772e-06,0.00171573377963928,signed,granite-3.3-8b-base
|
| 2899 |
xplain1.4-7base-32k-step1743,Layer 00 | self_attn.q_proj,2,Layer 00,0,self_attn.q_proj,2,16777216,-6.209917557953304e-08,0.0009247158851575594,signed,granite-3.3-8b-base
|
| 2900 |
xplain1.4-7base-32k-step1743,Layer 00 | self_attn.k_proj,3,Layer 00,0,self_attn.k_proj,3,4194304,8.879782171788975e-07,0.000822534841971489,signed,granite-3.3-8b-base
|
| 2901 |
xplain1.4-7base-32k-step1743,Layer 00 | self_attn.v_proj,4,Layer 00,0,self_attn.v_proj,4,4194304,-1.3729935233186552e-07,0.000704804584736643,signed,granite-3.3-8b-base
|
|
@@ -3256,4 +3239,3 @@ xplain1.4-7base-32k-step1743,Layer 39 | mlp.up_proj,39007,Layer 39,39,mlp.up_pro
|
|
| 3256 |
xplain1.4-7base-32k-step1743,Layer 39 | mlp.down_proj,39008,Layer 39,39,mlp.down_proj,8,52428800,-1.0088945600728039e-07,0.0012959232895043912,signed,granite-3.3-8b-base
|
| 3257 |
xplain1.4-7base-32k-step1743,Layer 39 | input_layernorm,39009,Layer 39,39,input_layernorm,9,4096,-0.006230114493519068,0.0015059144296379078,signed,granite-3.3-8b-base
|
| 3258 |
xplain1.4-7base-32k-step1743,Layer 39 | post_attention_layernorm,39010,Layer 39,39,post_attention_layernorm,10,4096,-0.000734880450181663,0.0016653910420154584,signed,granite-3.3-8b-base
|
| 3259 |
-
xplain1.4-7base-32k-step1743,Final Norm | model.norm.weight,900000,Final Norm,900000,final_norm,16,4096,0.012469591572880745,0.006104727263580456,signed,granite-3.3-8b-base
|
|
|
|
| 1 |
model,sublayer,sublayer_sort,block_label,block_sort,module_type,module_sort,count,mean,std,stat_mode,base_model
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
xplain1.3-6base-16k-step2199,Layer 00 | self_attn.q_proj,2,Layer 00,0,self_attn.q_proj,2,16777216,-2.5386498236912303e-08,0.001145262995459996,signed,granite-3.1-8b-base
|
| 3 |
xplain1.3-6base-32k-step1647,Layer 00 | self_attn.q_proj,2,Layer 00,0,self_attn.q_proj,2,16777216,2.6751570203487063e-07,0.001236013762140753,signed,granite-3.1-8b-base
|
| 4 |
xplain1.3-6base-64k-step1647,Layer 00 | self_attn.q_proj,2,Layer 00,0,self_attn.q_proj,2,16777216,3.582210297281563e-07,0.0014320992904867204,signed,granite-3.1-8b-base
|
|
|
|
| 2879 |
xplain1.4-6base-32k-step1743,Layer 39 | post_attention_layernorm,39010,Layer 39,39,post_attention_layernorm,10,4096,0.0011097422102466226,0.0020724020711700023,signed,granite-3.1-8b-base
|
| 2880 |
xplain1.4-6base-64k-step1743,Layer 39 | post_attention_layernorm,39010,Layer 39,39,post_attention_layernorm,10,4096,0.00044853618601337075,0.002316154528465667,signed,granite-3.1-8b-base
|
| 2881 |
xplain1.4-6base-8k-step2319,Layer 39 | post_attention_layernorm,39010,Layer 39,39,post_attention_layernorm,10,4096,0.0006851548096165061,0.0010263779705000592,signed,granite-3.1-8b-base
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2882 |
xplain1.4-7base-32k-step1743,Layer 00 | self_attn.q_proj,2,Layer 00,0,self_attn.q_proj,2,16777216,-6.209917557953304e-08,0.0009247158851575594,signed,granite-3.3-8b-base
|
| 2883 |
xplain1.4-7base-32k-step1743,Layer 00 | self_attn.k_proj,3,Layer 00,0,self_attn.k_proj,3,4194304,8.879782171788975e-07,0.000822534841971489,signed,granite-3.3-8b-base
|
| 2884 |
xplain1.4-7base-32k-step1743,Layer 00 | self_attn.v_proj,4,Layer 00,0,self_attn.v_proj,4,4194304,-1.3729935233186552e-07,0.000704804584736643,signed,granite-3.3-8b-base
|
|
|
|
| 3239 |
xplain1.4-7base-32k-step1743,Layer 39 | mlp.down_proj,39008,Layer 39,39,mlp.down_proj,8,52428800,-1.0088945600728039e-07,0.0012959232895043912,signed,granite-3.3-8b-base
|
| 3240 |
xplain1.4-7base-32k-step1743,Layer 39 | input_layernorm,39009,Layer 39,39,input_layernorm,9,4096,-0.006230114493519068,0.0015059144296379078,signed,granite-3.3-8b-base
|
| 3241 |
xplain1.4-7base-32k-step1743,Layer 39 | post_attention_layernorm,39010,Layer 39,39,post_attention_layernorm,10,4096,-0.000734880450181663,0.0016653910420154584,signed,granite-3.3-8b-base
|
|
|
sublayer_std_granite-3.1-8b-base.pdf
CHANGED
|
Binary files a/sublayer_std_granite-3.1-8b-base.pdf and b/sublayer_std_granite-3.1-8b-base.pdf differ
|
|
|
sublayer_std_granite-3.3-8b-base.pdf
CHANGED
|
Binary files a/sublayer_std_granite-3.3-8b-base.pdf and b/sublayer_std_granite-3.3-8b-base.pdf differ
|
|
|
summary.md
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|