SEWY2-640M-untrained / config.json
Aarushhh's picture
lower final logit softcapping from 30 to 15 ( as kellerjordans x post reccomends)
2e3abef verified
{
"architectures": [
"SewyV2ForCausalLM"
],
"auto_map": {
"AutoConfig": "modular_Sewy2.SewyV2Config",
"AutoModel": "modular_Sewy2.SewyV2Model",
"AutoModelForCausalLM": "modular_Sewy2.SewyV2ForCausalLM"
},
"attention_bias": false,
"attention_dropout": 0.0,
"attn_logit_softcapping": 50.0,
"aux_loss_alpha": 0.001,
"bos_token_id": 1,
"eos_token_id": 2,
"ep_size": 1,
"final_logit_softcapping": 15.0,
"first_k_dense_replace": 3,
"hidden_act": "relu2",
"hidden_size": 400,
"initializer_range": 0.02,
"intermediate_size": 1200,
"kv_lora_rank": 128,
"max_position_embeddings": 163840,
"model_type": "Sewy_v2",
"moe_intermediate_size": 200,
"moe_layer_freq": 1,
"n_group": 8,
"n_routed_experts": 64,
"n_shared_experts": 1,
"neutreno_lambda": 0.4,
"norm_topk_prob": true,
"num_attention_heads": 32,
"num_experts_per_tok": 8,
"num_hidden_layers": 32,
"num_key_value_heads": 32,
"num_nextn_predict_layers": 1,
"pretraining_tp": 1,
"q_lora_rank": 256,
"qk_nope_head_dim": 128,
"qk_rope_head_dim": 64,
"resformer_lambda": 2.0,
"rms_norm_eps": 1e-06,
"rope_scaling": {
"beta_fast": 32,
"beta_slow": 1,
"factor": 40,
"mscale": 1.0,
"mscale_all_dim": 1.0,
"original_max_position_embeddings": 4096,
"type": "yarn"
},
"rope_theta": 10000,
"routed_scaling_factor": 2.5,
"scoring_func": "sigmoid",
"seq_aux": true,
"tie_word_embeddings": false,
"topk_group": 4,
"topk_method": "noaux_tc",
"torch_dtype": "float32",
"transformers_version": "4.44.2",
"unit_norm_eps": 1e-06,
"use_cache": true,
"v_head_dim": 128,
"vocab_size": 49152
}