| { | |
| "model.decoder.layers.0.self_attn.k_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.0.self_attn.v_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.0.self_attn.q_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.0.self_attn.out_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.0.encoder_attn.k_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.0.encoder_attn.v_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.0.encoder_attn.q_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.0.encoder_attn.out_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.0.fc1": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.0.fc2": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.1.self_attn.k_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.1.self_attn.v_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.1.self_attn.q_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.1.self_attn.out_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.1.encoder_attn.k_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.1.encoder_attn.v_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.1.encoder_attn.q_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.1.encoder_attn.out_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.1.fc1": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.1.fc2": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.2.self_attn.k_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.2.self_attn.v_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.2.self_attn.q_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.2.self_attn.out_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.2.encoder_attn.k_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.2.encoder_attn.v_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.2.encoder_attn.q_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.2.encoder_attn.out_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.2.fc1": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.2.fc2": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.3.self_attn.k_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.3.self_attn.v_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.3.self_attn.q_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.3.self_attn.out_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.3.encoder_attn.k_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.3.encoder_attn.v_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.3.encoder_attn.q_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.3.encoder_attn.out_proj": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.3.fc1": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "model.decoder.layers.3.fc2": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| }, | |
| "proj_out": { | |
| "weights": "qint8", | |
| "activations": "none" | |
| } | |
| } |