mirror of
https://github.com/KoboldAI/KoboldAI-Client.git
synced 2024-12-12 08:36:28 +01:00
33 lines
2.3 KiB
JSON
33 lines
2.3 KiB
JSON
{
|
|
"mtj_compat": "neo",
|
|
"mtj_pe": "fixed",
|
|
"mtj_config_map": {
|
|
"d_model": "hidden_size",
|
|
"n_heads": "num_heads",
|
|
"layers": "num_layers"
|
|
},
|
|
"static_weights": {
|
|
"transformer.wte.weight": {"mtj": {"module": "embedding_shard/~/linear", "param": "w", "transforms": ["no_transpose", "vocab_pad"]}},
|
|
"transformer.wpe.weight": {"mtj": {"module": "embedding_shard", "param": "pos_embs", "transforms": ["no_transpose"]}},
|
|
"transformer.ln_f.weight": {"mtj": {"module": "projection_shard/~/replicated_layer_norm", "param": "scale"}},
|
|
"transformer.ln_f.bias": {"mtj": {"module": "projection_shard/~/replicated_layer_norm", "param": "offset"}}
|
|
},
|
|
"layer_weights": {
|
|
"transformer.h.{layer}.attn.attention.bias": {},
|
|
"transformer.h.{layer}.attn.attention.masked_bias": {},
|
|
"transformer.h.{layer}.attn.attention.q_proj.weight": {"mtj": {"module": "layer_{layer}/~/linear", "param": "w"}},
|
|
"transformer.h.{layer}.attn.attention.v_proj.weight": {"mtj": {"module": "layer_{layer}/~/linear_1", "param": "w"}},
|
|
"transformer.h.{layer}.attn.attention.k_proj.weight": {"mtj": {"module": "layer_{layer}/~/linear_2", "param": "w"}},
|
|
"transformer.h.{layer}.attn.attention.out_proj.weight": {"mtj": {"module": "layer_{layer}/~/linear_3", "param": "w"}},
|
|
"transformer.h.{layer}.attn.attention.out_proj.bias": {"mtj": {"module": "layer_{layer}/~/linear_3", "param": "b", "transforms": ["divide_by_shards"]}},
|
|
"transformer.h.{layer}.mlp.c_fc.weight": {"mtj": {"module": "layer_{layer}/~/linear_4", "param": "w"}},
|
|
"transformer.h.{layer}.mlp.c_fc.bias": {"mtj": {"module": "layer_{layer}/~/linear_4", "param": "b"}},
|
|
"transformer.h.{layer}.mlp.c_proj.weight": {"mtj": {"module": "layer_{layer}/~/linear_5", "param": "w"}},
|
|
"transformer.h.{layer}.mlp.c_proj.bias": {"mtj": {"module": "layer_{layer}/~/linear_5", "param": "b", "transforms": ["divide_by_shards"]}},
|
|
"transformer.h.{layer}.ln_1.weight": {"mtj": {"module": "layer_{layer}/~/replicated_layer_norm", "param": "scale"}},
|
|
"transformer.h.{layer}.ln_1.bias": {"mtj": {"module": "layer_{layer}/~/replicated_layer_norm", "param": "offset"}},
|
|
"transformer.h.{layer}.ln_2.weight": {"mtj": {"module": "layer_{layer}/~/replicated_layer_norm_1", "param": "scale"}},
|
|
"transformer.h.{layer}.ln_2.bias": {"mtj": {"module": "layer_{layer}/~/replicated_layer_norm_1", "param": "offset"}}
|
|
}
|
|
}
|