Skip to content

Commit 3dbb1ec

Browse files
xaduprexadupre
committed
add cached config
1 parent b9fa045 commit 3dbb1ec

File tree

1 file changed

+142
-0
lines changed

1 file changed

+142
-0
lines changed

‎onnx_diagnostic/torch_models/hghub/hub_data_cached_configs.py‎

Lines changed: 142 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -4160,3 +4160,145 @@ def _ccached_ydshieh_tiny_random_vit_for_image_classification():
41604160
"transformers_version": "4.24.0.dev0",
41614161
}
41624162
)
4163+
4164+
4165+
def _ccached_microsoft_phi_35_mini_instruct():
4166+
"microsoft/Phi-3.5-mini-instruct"
4167+
return transformers.Phi3Config(
4168+
**{
4169+
"_name_or_path": "Phi-3.5-mini-instruct",
4170+
"architectures": ["Phi3ForCausalLM"],
4171+
"attention_dropout": 0.0,
4172+
"auto_map": {
4173+
"AutoConfig": "configuration_phi3.Phi3Config",
4174+
"AutoModelForCausalLM": "modeling_phi3.Phi3ForCausalLM",
4175+
},
4176+
"bos_token_id": 1,
4177+
"embd_pdrop": 0.0,
4178+
"eos_token_id": 32000,
4179+
"hidden_act": "silu",
4180+
"hidden_size": 3072,
4181+
"initializer_range": 0.02,
4182+
"intermediate_size": 8192,
4183+
"max_position_embeddings": 131072,
4184+
"model_type": "phi3",
4185+
"num_attention_heads": 32,
4186+
"num_hidden_layers": 32,
4187+
"num_key_value_heads": 32,
4188+
"original_max_position_embeddings": 4096,
4189+
"pad_token_id": 32000,
4190+
"resid_pdrop": 0.0,
4191+
"rms_norm_eps": 1e-05,
4192+
"rope_scaling": {
4193+
"long_factor": [
4194+
1.0800000429153442,
4195+
1.1100000143051147,
4196+
1.1399999856948853,
4197+
1.340000033378601,
4198+
1.5899999141693115,
4199+
1.600000023841858,
4200+
1.6200000047683716,
4201+
2.620000123977661,
4202+
3.2300000190734863,
4203+
3.2300000190734863,
4204+
4.789999961853027,
4205+
7.400000095367432,
4206+
7.700000286102295,
4207+
9.09000015258789,
4208+
12.199999809265137,
4209+
17.670000076293945,
4210+
24.46000099182129,
4211+
28.57000160217285,
4212+
30.420001983642578,
4213+
30.840002059936523,
4214+
32.590003967285156,
4215+
32.93000411987305,
4216+
42.320003509521484,
4217+
44.96000289916992,
4218+
50.340003967285156,
4219+
50.45000457763672,
4220+
57.55000305175781,
4221+
57.93000411987305,
4222+
58.21000289916992,
4223+
60.1400032043457,
4224+
62.61000442504883,
4225+
62.62000274658203,
4226+
62.71000289916992,
4227+
63.1400032043457,
4228+
63.1400032043457,
4229+
63.77000427246094,
4230+
63.93000411987305,
4231+
63.96000289916992,
4232+
63.970001220703125,
4233+
64.02999877929688,
4234+
64.06999969482422,
4235+
64.08000183105469,
4236+
64.12000274658203,
4237+
64.41000366210938,
4238+
64.4800033569336,
4239+
64.51000213623047,
4240+
64.52999877929688,
4241+
64.83999633789062,
4242+
],
4243+
"short_factor": [
4244+
1.0,
4245+
1.0199999809265137,
4246+
1.0299999713897705,
4247+
1.0299999713897705,
4248+
1.0499999523162842,
4249+
1.0499999523162842,
4250+
1.0499999523162842,
4251+
1.0499999523162842,
4252+
1.0499999523162842,
4253+
1.0699999332427979,
4254+
1.0999999046325684,
4255+
1.1099998950958252,
4256+
1.1599998474121094,
4257+
1.1599998474121094,
4258+
1.1699998378753662,
4259+
1.2899998426437378,
4260+
1.339999794960022,
4261+
1.679999828338623,
4262+
1.7899998426437378,
4263+
1.8199998140335083,
4264+
1.8499997854232788,
4265+
1.8799997568130493,
4266+
1.9099997282028198,
4267+
1.9399996995925903,
4268+
1.9899996519088745,
4269+
2.0199997425079346,
4270+
2.0199997425079346,
4271+
2.0199997425079346,
4272+
2.0199997425079346,
4273+
2.0199997425079346,
4274+
2.0199997425079346,
4275+
2.0299997329711914,
4276+
2.0299997329711914,
4277+
2.0299997329711914,
4278+
2.0299997329711914,
4279+
2.0299997329711914,
4280+
2.0299997329711914,
4281+
2.0299997329711914,
4282+
2.0299997329711914,
4283+
2.0299997329711914,
4284+
2.0799996852874756,
4285+
2.0899996757507324,
4286+
2.189999580383301,
4287+
2.2199995517730713,
4288+
2.5899994373321533,
4289+
2.729999542236328,
4290+
2.749999523162842,
4291+
2.8399994373321533,
4292+
],
4293+
"type": "longrope",
4294+
},
4295+
"rope_theta": 10000.0,
4296+
"sliding_window": 262144,
4297+
"tie_word_embeddings": false,
4298+
"torch_dtype": "bfloat16",
4299+
"transformers_version": "4.43.3",
4300+
"use_cache": true,
4301+
"attention_bias": false,
4302+
"vocab_size": 32064,
4303+
}
4304+
)

0 commit comments

Comments
 (0)