anonym-repos's picture
Upload T5ForConditionalGeneration
202d893 verified
raw
history blame contribute delete
No virus
21.8 kB
{
"alpha_pattern": {},
"auto_mapping": null,
"base_model_name_or_path": "MU-NLPC/calcformer-instruct-flan-xl_step-128k",
"bias": "none",
"fan_in_fan_out": false,
"inference_mode": true,
"init_lora_weights": true,
"layers_pattern": null,
"layers_to_transform": null,
"loftq_config": {},
"lora_alpha": 32,
"lora_dropout": 0.1,
"megatron_config": null,
"megatron_core": "megatron.core",
"modules_to_save": null,
"peft_type": "LORA",
"r": 32,
"rank_pattern": {},
"revision": null,
"target_modules": [
"encoder.block.6.layer.1.DenseReluDense.wi_0",
"decoder.block.12.layer.0.SelfAttention.o",
"decoder.block.4.layer.0.SelfAttention.k",
"decoder.block.9.layer.1.EncDecAttention.o",
"decoder.block.4.layer.0.SelfAttention.v",
"encoder.block.3.layer.0.SelfAttention.o",
"encoder.block.4.layer.1.DenseReluDense.wi_0",
"decoder.block.20.layer.1.EncDecAttention.k",
"encoder.block.7.layer.0.SelfAttention.v",
"decoder.block.15.layer.2.DenseReluDense.wi_1",
"encoder.block.15.layer.0.SelfAttention.o",
"decoder.block.17.layer.0.SelfAttention.v",
"encoder.block.12.layer.1.DenseReluDense.wi_1",
"encoder.block.3.layer.1.DenseReluDense.wi_1",
"encoder.block.22.layer.1.DenseReluDense.wi_1",
"decoder.block.9.layer.2.DenseReluDense.wo",
"decoder.block.12.layer.0.SelfAttention.q",
"encoder.block.2.layer.1.DenseReluDense.wi_0",
"encoder.block.17.layer.0.SelfAttention.k",
"encoder.block.22.layer.0.SelfAttention.v",
"decoder.block.22.layer.1.EncDecAttention.k",
"decoder.block.16.layer.0.SelfAttention.o",
"decoder.block.3.layer.1.EncDecAttention.k",
"decoder.block.10.layer.2.DenseReluDense.wi_1",
"decoder.block.15.layer.0.SelfAttention.q",
"decoder.block.19.layer.1.EncDecAttention.q",
"encoder.block.3.layer.0.SelfAttention.v",
"decoder.block.22.layer.1.EncDecAttention.q",
"decoder.block.22.layer.1.EncDecAttention.o",
"decoder.block.5.layer.0.SelfAttention.o",
"encoder.block.11.layer.1.DenseReluDense.wi_1",
"decoder.block.13.layer.2.DenseReluDense.wi_0",
"decoder.block.14.layer.0.SelfAttention.o",
"encoder.block.6.layer.1.DenseReluDense.wo",
"decoder.block.17.layer.2.DenseReluDense.wi_1",
"decoder.block.0.layer.1.EncDecAttention.q",
"decoder.block.6.layer.1.EncDecAttention.v",
"decoder.block.3.layer.0.SelfAttention.k",
"encoder.block.5.layer.0.SelfAttention.o",
"encoder.block.8.layer.0.SelfAttention.v",
"decoder.block.19.layer.2.DenseReluDense.wi_1",
"decoder.block.22.layer.0.SelfAttention.o",
"encoder.block.11.layer.0.SelfAttention.k",
"decoder.block.10.layer.2.DenseReluDense.wi_0",
"decoder.block.11.layer.0.SelfAttention.v",
"decoder.block.7.layer.1.EncDecAttention.k",
"decoder.block.21.layer.0.SelfAttention.k",
"decoder.block.3.layer.1.EncDecAttention.q",
"decoder.block.20.layer.1.EncDecAttention.v",
"encoder.block.9.layer.0.SelfAttention.o",
"decoder.block.11.layer.1.EncDecAttention.v",
"encoder.block.15.layer.0.SelfAttention.q",
"encoder.block.9.layer.1.DenseReluDense.wi_1",
"encoder.block.20.layer.0.SelfAttention.o",
"decoder.block.19.layer.1.EncDecAttention.o",
"decoder.block.0.layer.0.SelfAttention.v",
"decoder.block.19.layer.0.SelfAttention.o",
"decoder.block.23.layer.0.SelfAttention.q",
"encoder.block.16.layer.0.SelfAttention.q",
"decoder.block.6.layer.2.DenseReluDense.wo",
"decoder.block.20.layer.1.EncDecAttention.o",
"decoder.block.15.layer.0.SelfAttention.v",
"decoder.block.16.layer.1.EncDecAttention.v",
"decoder.block.0.layer.0.SelfAttention.q",
"decoder.block.21.layer.2.DenseReluDense.wi_1",
"encoder.block.3.layer.1.DenseReluDense.wi_0",
"encoder.block.5.layer.0.SelfAttention.q",
"decoder.block.15.layer.0.SelfAttention.o",
"encoder.block.18.layer.0.SelfAttention.o",
"decoder.block.21.layer.1.EncDecAttention.q",
"encoder.block.18.layer.0.SelfAttention.q",
"decoder.block.19.layer.0.SelfAttention.k",
"encoder.block.16.layer.1.DenseReluDense.wo",
"decoder.block.9.layer.1.EncDecAttention.k",
"decoder.block.6.layer.0.SelfAttention.q",
"decoder.block.0.layer.1.EncDecAttention.o",
"decoder.block.17.layer.2.DenseReluDense.wi_0",
"decoder.block.16.layer.0.SelfAttention.v",
"decoder.block.17.layer.0.SelfAttention.o",
"decoder.block.20.layer.0.SelfAttention.k",
"encoder.block.23.layer.1.DenseReluDense.wi_0",
"encoder.block.2.layer.1.DenseReluDense.wo",
"decoder.block.1.layer.1.EncDecAttention.o",
"decoder.block.15.layer.1.EncDecAttention.q",
"decoder.block.2.layer.2.DenseReluDense.wi_1",
"decoder.block.22.layer.2.DenseReluDense.wi_0",
"encoder.block.16.layer.0.SelfAttention.k",
"decoder.block.16.layer.2.DenseReluDense.wi_1",
"decoder.block.23.layer.1.EncDecAttention.q",
"decoder.block.1.layer.2.DenseReluDense.wo",
"decoder.block.23.layer.1.EncDecAttention.o",
"decoder.block.1.layer.1.EncDecAttention.q",
"decoder.block.17.layer.2.DenseReluDense.wo",
"decoder.block.13.layer.0.SelfAttention.q",
"decoder.block.3.layer.2.DenseReluDense.wi_1",
"decoder.block.16.layer.2.DenseReluDense.wi_0",
"decoder.block.21.layer.0.SelfAttention.q",
"decoder.block.18.layer.0.SelfAttention.o",
"decoder.block.3.layer.1.EncDecAttention.o",
"decoder.block.2.layer.1.EncDecAttention.o",
"encoder.block.6.layer.0.SelfAttention.v",
"decoder.block.20.layer.2.DenseReluDense.wi_0",
"decoder.block.3.layer.1.EncDecAttention.v",
"encoder.block.2.layer.1.DenseReluDense.wi_1",
"decoder.block.11.layer.0.SelfAttention.o",
"decoder.block.22.layer.2.DenseReluDense.wi_1",
"decoder.block.22.layer.1.EncDecAttention.v",
"encoder.block.22.layer.1.DenseReluDense.wi_0",
"decoder.block.18.layer.2.DenseReluDense.wi_1",
"encoder.block.19.layer.0.SelfAttention.q",
"encoder.block.21.layer.0.SelfAttention.k",
"decoder.block.10.layer.0.SelfAttention.q",
"encoder.block.11.layer.0.SelfAttention.o",
"encoder.block.17.layer.1.DenseReluDense.wi_0",
"decoder.block.18.layer.1.EncDecAttention.k",
"decoder.block.4.layer.1.EncDecAttention.q",
"decoder.block.7.layer.1.EncDecAttention.q",
"decoder.block.19.layer.0.SelfAttention.q",
"decoder.block.12.layer.1.EncDecAttention.v",
"decoder.block.3.layer.0.SelfAttention.v",
"decoder.block.8.layer.0.SelfAttention.k",
"encoder.block.2.layer.0.SelfAttention.k",
"decoder.block.10.layer.0.SelfAttention.v",
"decoder.block.5.layer.2.DenseReluDense.wo",
"decoder.block.14.layer.1.EncDecAttention.q",
"encoder.block.6.layer.0.SelfAttention.k",
"encoder.block.12.layer.1.DenseReluDense.wo",
"decoder.block.11.layer.2.DenseReluDense.wi_0",
"decoder.block.15.layer.1.EncDecAttention.v",
"decoder.block.14.layer.0.SelfAttention.v",
"decoder.block.23.layer.2.DenseReluDense.wo",
"decoder.block.1.layer.2.DenseReluDense.wi_1",
"decoder.block.6.layer.0.SelfAttention.o",
"encoder.block.0.layer.1.DenseReluDense.wo",
"decoder.block.18.layer.1.EncDecAttention.q",
"decoder.block.17.layer.1.EncDecAttention.v",
"decoder.block.10.layer.1.EncDecAttention.o",
"decoder.block.11.layer.1.EncDecAttention.o",
"decoder.block.19.layer.0.SelfAttention.v",
"decoder.block.13.layer.1.EncDecAttention.q",
"decoder.block.16.layer.0.SelfAttention.k",
"decoder.block.9.layer.1.EncDecAttention.q",
"encoder.block.9.layer.0.SelfAttention.k",
"decoder.block.5.layer.0.SelfAttention.v",
"decoder.block.4.layer.0.SelfAttention.o",
"decoder.block.10.layer.0.SelfAttention.k",
"encoder.block.3.layer.0.SelfAttention.k",
"encoder.block.13.layer.0.SelfAttention.k",
"decoder.block.14.layer.0.SelfAttention.q",
"encoder.block.17.layer.1.DenseReluDense.wi_1",
"decoder.block.0.layer.2.DenseReluDense.wo",
"encoder.block.16.layer.0.SelfAttention.o",
"decoder.block.0.layer.2.DenseReluDense.wi_0",
"encoder.block.17.layer.0.SelfAttention.v",
"encoder.block.19.layer.0.SelfAttention.o",
"decoder.block.8.layer.2.DenseReluDense.wo",
"decoder.block.22.layer.0.SelfAttention.k",
"encoder.block.13.layer.1.DenseReluDense.wi_1",
"encoder.block.16.layer.1.DenseReluDense.wi_1",
"decoder.block.9.layer.2.DenseReluDense.wi_1",
"decoder.block.2.layer.0.SelfAttention.k",
"encoder.block.17.layer.0.SelfAttention.q",
"decoder.block.16.layer.2.DenseReluDense.wo",
"decoder.block.6.layer.2.DenseReluDense.wi_0",
"decoder.block.4.layer.2.DenseReluDense.wi_0",
"decoder.block.11.layer.1.EncDecAttention.k",
"decoder.block.2.layer.1.EncDecAttention.q",
"encoder.block.5.layer.1.DenseReluDense.wo",
"decoder.block.2.layer.2.DenseReluDense.wo",
"encoder.block.10.layer.1.DenseReluDense.wi_1",
"decoder.block.18.layer.2.DenseReluDense.wi_0",
"encoder.block.10.layer.0.SelfAttention.k",
"encoder.block.21.layer.1.DenseReluDense.wo",
"decoder.block.8.layer.0.SelfAttention.o",
"decoder.block.10.layer.2.DenseReluDense.wo",
"encoder.block.11.layer.0.SelfAttention.v",
"decoder.block.5.layer.0.SelfAttention.q",
"decoder.block.19.layer.2.DenseReluDense.wi_0",
"decoder.block.7.layer.0.SelfAttention.k",
"decoder.block.8.layer.1.EncDecAttention.o",
"encoder.block.12.layer.0.SelfAttention.o",
"encoder.block.0.layer.1.DenseReluDense.wi_1",
"encoder.block.3.layer.0.SelfAttention.q",
"decoder.block.1.layer.0.SelfAttention.v",
"encoder.block.18.layer.0.SelfAttention.v",
"encoder.block.19.layer.1.DenseReluDense.wo",
"decoder.block.23.layer.1.EncDecAttention.v",
"encoder.block.14.layer.0.SelfAttention.k",
"decoder.block.12.layer.2.DenseReluDense.wi_0",
"decoder.block.20.layer.0.SelfAttention.v",
"decoder.block.14.layer.1.EncDecAttention.o",
"decoder.block.11.layer.0.SelfAttention.q",
"encoder.block.15.layer.0.SelfAttention.k",
"decoder.block.23.layer.2.DenseReluDense.wi_0",
"decoder.block.2.layer.0.SelfAttention.o",
"decoder.block.0.layer.2.DenseReluDense.wi_1",
"encoder.block.14.layer.0.SelfAttention.q",
"encoder.block.12.layer.0.SelfAttention.v",
"encoder.block.20.layer.1.DenseReluDense.wi_0",
"encoder.block.19.layer.1.DenseReluDense.wi_1",
"decoder.block.6.layer.2.DenseReluDense.wi_1",
"decoder.block.7.layer.0.SelfAttention.v",
"encoder.block.7.layer.1.DenseReluDense.wo",
"decoder.block.4.layer.1.EncDecAttention.k",
"decoder.block.1.layer.1.EncDecAttention.k",
"decoder.block.17.layer.0.SelfAttention.q",
"encoder.block.9.layer.0.SelfAttention.v",
"encoder.block.19.layer.0.SelfAttention.k",
"encoder.block.1.layer.0.SelfAttention.k",
"decoder.block.16.layer.1.EncDecAttention.q",
"decoder.block.18.layer.1.EncDecAttention.o",
"encoder.block.20.layer.0.SelfAttention.q",
"decoder.block.14.layer.2.DenseReluDense.wi_0",
"encoder.block.23.layer.0.SelfAttention.k",
"decoder.block.21.layer.1.EncDecAttention.k",
"encoder.block.22.layer.1.DenseReluDense.wo",
"decoder.block.13.layer.1.EncDecAttention.k",
"decoder.block.10.layer.1.EncDecAttention.k",
"encoder.block.20.layer.1.DenseReluDense.wo",
"decoder.block.2.layer.1.EncDecAttention.v",
"encoder.block.23.layer.0.SelfAttention.v",
"decoder.block.14.layer.2.DenseReluDense.wo",
"encoder.block.5.layer.0.SelfAttention.v",
"decoder.block.10.layer.1.EncDecAttention.v",
"encoder.block.14.layer.1.DenseReluDense.wi_0",
"encoder.block.13.layer.1.DenseReluDense.wo",
"encoder.block.8.layer.1.DenseReluDense.wi_0",
"encoder.block.3.layer.1.DenseReluDense.wo",
"encoder.block.5.layer.1.DenseReluDense.wi_1",
"decoder.block.6.layer.1.EncDecAttention.q",
"encoder.block.14.layer.0.SelfAttention.o",
"decoder.block.23.layer.0.SelfAttention.v",
"encoder.block.18.layer.1.DenseReluDense.wi_1",
"decoder.block.7.layer.2.DenseReluDense.wi_0",
"decoder.block.17.layer.1.EncDecAttention.q",
"decoder.block.20.layer.0.SelfAttention.q",
"encoder.block.12.layer.0.SelfAttention.q",
"encoder.block.8.layer.0.SelfAttention.q",
"encoder.block.6.layer.1.DenseReluDense.wi_1",
"decoder.block.4.layer.2.DenseReluDense.wi_1",
"encoder.block.18.layer.0.SelfAttention.k",
"decoder.block.15.layer.1.EncDecAttention.k",
"decoder.block.14.layer.0.SelfAttention.k",
"decoder.block.2.layer.0.SelfAttention.q",
"decoder.block.15.layer.2.DenseReluDense.wo",
"encoder.block.14.layer.0.SelfAttention.v",
"decoder.block.13.layer.1.EncDecAttention.v",
"encoder.block.20.layer.0.SelfAttention.v",
"decoder.block.12.layer.0.SelfAttention.k",
"encoder.block.19.layer.0.SelfAttention.v",
"decoder.block.11.layer.2.DenseReluDense.wo",
"encoder.block.1.layer.1.DenseReluDense.wi_0",
"encoder.block.16.layer.0.SelfAttention.v",
"encoder.block.1.layer.0.SelfAttention.v",
"decoder.block.19.layer.1.EncDecAttention.v",
"decoder.block.3.layer.0.SelfAttention.q",
"decoder.block.7.layer.2.DenseReluDense.wo",
"decoder.block.22.layer.0.SelfAttention.q",
"decoder.block.8.layer.2.DenseReluDense.wi_1",
"encoder.block.1.layer.1.DenseReluDense.wo",
"decoder.block.3.layer.0.SelfAttention.o",
"encoder.block.13.layer.1.DenseReluDense.wi_0",
"decoder.block.9.layer.0.SelfAttention.k",
"decoder.block.7.layer.0.SelfAttention.o",
"decoder.block.2.layer.0.SelfAttention.v",
"decoder.block.13.layer.0.SelfAttention.v",
"encoder.block.22.layer.0.SelfAttention.q",
"decoder.block.9.layer.1.EncDecAttention.v",
"encoder.block.13.layer.0.SelfAttention.v",
"decoder.block.13.layer.2.DenseReluDense.wo",
"encoder.block.11.layer.1.DenseReluDense.wi_0",
"encoder.block.7.layer.0.SelfAttention.k",
"decoder.block.5.layer.1.EncDecAttention.o",
"decoder.block.9.layer.2.DenseReluDense.wi_0",
"encoder.block.13.layer.0.SelfAttention.o",
"decoder.block.18.layer.0.SelfAttention.v",
"encoder.block.18.layer.1.DenseReluDense.wi_0",
"encoder.block.11.layer.0.SelfAttention.q",
"encoder.block.2.layer.0.SelfAttention.o",
"decoder.block.8.layer.1.EncDecAttention.v",
"encoder.block.0.layer.1.DenseReluDense.wi_0",
"decoder.block.12.layer.0.SelfAttention.v",
"decoder.block.9.layer.0.SelfAttention.o",
"encoder.block.1.layer.1.DenseReluDense.wi_1",
"decoder.block.21.layer.1.EncDecAttention.v",
"decoder.block.20.layer.2.DenseReluDense.wo",
"encoder.block.5.layer.1.DenseReluDense.wi_0",
"encoder.block.18.layer.1.DenseReluDense.wo",
"encoder.block.8.layer.0.SelfAttention.o",
"decoder.block.17.layer.1.EncDecAttention.k",
"encoder.block.1.layer.0.SelfAttention.q",
"encoder.block.7.layer.1.DenseReluDense.wi_1",
"decoder.block.23.layer.1.EncDecAttention.k",
"decoder.block.16.layer.0.SelfAttention.q",
"decoder.block.7.layer.2.DenseReluDense.wi_1",
"encoder.block.9.layer.1.DenseReluDense.wo",
"encoder.block.23.layer.1.DenseReluDense.wi_1",
"decoder.block.0.layer.0.SelfAttention.k",
"decoder.block.7.layer.0.SelfAttention.q",
"encoder.block.11.layer.1.DenseReluDense.wo",
"decoder.block.13.layer.0.SelfAttention.o",
"decoder.block.9.layer.0.SelfAttention.q",
"encoder.block.9.layer.1.DenseReluDense.wi_0",
"decoder.block.22.layer.2.DenseReluDense.wo",
"encoder.block.1.layer.0.SelfAttention.o",
"encoder.block.0.layer.0.SelfAttention.k",
"encoder.block.5.layer.0.SelfAttention.k",
"decoder.block.12.layer.2.DenseReluDense.wi_1",
"encoder.block.21.layer.0.SelfAttention.v",
"decoder.block.21.layer.2.DenseReluDense.wo",
"decoder.block.17.layer.0.SelfAttention.k",
"decoder.block.0.layer.1.EncDecAttention.k",
"encoder.block.0.layer.0.SelfAttention.q",
"encoder.block.9.layer.0.SelfAttention.q",
"encoder.block.21.layer.0.SelfAttention.q",
"decoder.block.2.layer.2.DenseReluDense.wi_0",
"encoder.block.4.layer.0.SelfAttention.k",
"encoder.block.10.layer.0.SelfAttention.o",
"encoder.block.14.layer.1.DenseReluDense.wi_1",
"decoder.block.8.layer.1.EncDecAttention.q",
"decoder.block.6.layer.1.EncDecAttention.k",
"decoder.block.20.layer.2.DenseReluDense.wi_1",
"encoder.block.19.layer.1.DenseReluDense.wi_0",
"decoder.block.6.layer.0.SelfAttention.v",
"decoder.block.8.layer.2.DenseReluDense.wi_0",
"decoder.block.14.layer.2.DenseReluDense.wi_1",
"encoder.block.15.layer.1.DenseReluDense.wo",
"decoder.block.20.layer.1.EncDecAttention.q",
"decoder.block.11.layer.2.DenseReluDense.wi_1",
"encoder.block.4.layer.1.DenseReluDense.wo",
"encoder.block.10.layer.0.SelfAttention.v",
"decoder.block.18.layer.2.DenseReluDense.wo",
"encoder.block.8.layer.1.DenseReluDense.wi_1",
"encoder.block.10.layer.0.SelfAttention.q",
"encoder.block.2.layer.0.SelfAttention.v",
"encoder.block.0.layer.0.SelfAttention.v",
"decoder.block.4.layer.2.DenseReluDense.wo",
"encoder.block.17.layer.1.DenseReluDense.wo",
"decoder.block.7.layer.1.EncDecAttention.v",
"encoder.block.15.layer.1.DenseReluDense.wi_1",
"encoder.block.10.layer.1.DenseReluDense.wo",
"decoder.block.12.layer.1.EncDecAttention.k",
"encoder.block.4.layer.1.DenseReluDense.wi_1",
"decoder.block.8.layer.0.SelfAttention.q",
"decoder.block.20.layer.0.SelfAttention.o",
"decoder.block.22.layer.0.SelfAttention.v",
"decoder.block.1.layer.0.SelfAttention.k",
"encoder.block.15.layer.1.DenseReluDense.wi_0",
"decoder.block.4.layer.0.SelfAttention.q",
"encoder.block.2.layer.0.SelfAttention.q",
"decoder.block.10.layer.1.EncDecAttention.q",
"decoder.block.6.layer.1.EncDecAttention.o",
"decoder.block.11.layer.1.EncDecAttention.q",
"decoder.block.23.layer.0.SelfAttention.o",
"encoder.block.7.layer.0.SelfAttention.q",
"decoder.block.12.layer.1.EncDecAttention.o",
"encoder.block.12.layer.0.SelfAttention.k",
"encoder.block.22.layer.0.SelfAttention.o",
"encoder.block.21.layer.1.DenseReluDense.wi_1",
"decoder.block.8.layer.1.EncDecAttention.k",
"decoder.block.5.layer.0.SelfAttention.k",
"decoder.block.18.layer.0.SelfAttention.q",
"decoder.block.10.layer.0.SelfAttention.o",
"decoder.block.15.layer.0.SelfAttention.k",
"decoder.block.23.layer.0.SelfAttention.k",
"encoder.block.4.layer.0.SelfAttention.v",
"decoder.block.9.layer.0.SelfAttention.v",
"decoder.block.3.layer.2.DenseReluDense.wi_0",
"decoder.block.11.layer.0.SelfAttention.k",
"decoder.block.21.layer.0.SelfAttention.o",
"encoder.block.22.layer.0.SelfAttention.k",
"decoder.block.15.layer.1.EncDecAttention.o",
"decoder.block.4.layer.1.EncDecAttention.o",
"encoder.block.20.layer.0.SelfAttention.k",
"decoder.block.14.layer.1.EncDecAttention.v",
"decoder.block.0.layer.0.SelfAttention.o",
"encoder.block.21.layer.1.DenseReluDense.wi_0",
"encoder.block.8.layer.0.SelfAttention.k",
"decoder.block.16.layer.1.EncDecAttention.k",
"decoder.block.21.layer.0.SelfAttention.v",
"decoder.block.1.layer.0.SelfAttention.o",
"encoder.block.14.layer.1.DenseReluDense.wo",
"decoder.block.2.layer.1.EncDecAttention.k",
"decoder.block.18.layer.1.EncDecAttention.v",
"decoder.block.5.layer.2.DenseReluDense.wi_1",
"decoder.block.13.layer.1.EncDecAttention.o",
"encoder.block.6.layer.0.SelfAttention.q",
"decoder.block.23.layer.2.DenseReluDense.wi_1",
"encoder.block.23.layer.0.SelfAttention.q",
"encoder.block.12.layer.1.DenseReluDense.wi_0",
"decoder.block.5.layer.1.EncDecAttention.k",
"decoder.block.13.layer.2.DenseReluDense.wi_1",
"encoder.block.7.layer.0.SelfAttention.o",
"decoder.block.3.layer.2.DenseReluDense.wo",
"decoder.block.13.layer.0.SelfAttention.k",
"decoder.block.1.layer.1.EncDecAttention.v",
"encoder.block.20.layer.1.DenseReluDense.wi_1",
"lm_head",
"decoder.block.1.layer.2.DenseReluDense.wi_0",
"decoder.block.14.layer.1.EncDecAttention.k",
"encoder.block.21.layer.0.SelfAttention.o",
"encoder.block.4.layer.0.SelfAttention.q",
"decoder.block.5.layer.1.EncDecAttention.v",
"decoder.block.21.layer.2.DenseReluDense.wi_0",
"decoder.block.5.layer.2.DenseReluDense.wi_0",
"decoder.block.16.layer.1.EncDecAttention.o",
"decoder.block.18.layer.0.SelfAttention.k",
"decoder.block.5.layer.1.EncDecAttention.q",
"encoder.block.13.layer.0.SelfAttention.q",
"decoder.block.12.layer.1.EncDecAttention.q",
"decoder.block.19.layer.2.DenseReluDense.wo",
"encoder.block.4.layer.0.SelfAttention.o",
"decoder.block.8.layer.0.SelfAttention.v",
"encoder.block.10.layer.1.DenseReluDense.wi_0",
"decoder.block.0.layer.1.EncDecAttention.v",
"decoder.block.19.layer.1.EncDecAttention.k",
"encoder.block.23.layer.0.SelfAttention.o",
"decoder.block.6.layer.0.SelfAttention.k",
"encoder.block.15.layer.0.SelfAttention.v",
"decoder.block.7.layer.1.EncDecAttention.o",
"encoder.block.8.layer.1.DenseReluDense.wo",
"decoder.block.15.layer.2.DenseReluDense.wi_0",
"encoder.block.16.layer.1.DenseReluDense.wi_0",
"decoder.block.21.layer.1.EncDecAttention.o",
"encoder.block.7.layer.1.DenseReluDense.wi_0",
"decoder.block.12.layer.2.DenseReluDense.wo",
"decoder.block.4.layer.1.EncDecAttention.v",
"decoder.block.17.layer.1.EncDecAttention.o",
"encoder.block.0.layer.0.SelfAttention.o",
"decoder.block.1.layer.0.SelfAttention.q",
"encoder.block.23.layer.1.DenseReluDense.wo",
"encoder.block.17.layer.0.SelfAttention.o",
"encoder.block.6.layer.0.SelfAttention.o"
],
"task_type": "SEQ_2_SEQ_LM",
"use_rslora": false
}