gsa-1.3B-100B / config.json
yzhangcs's picture
Upload GSAForCausalLM
023f4e2 verified
raw
history blame
No virus
1 kB
{
"_name_or_path": "fla-hub/gsa-1.3B-100B",
"architectures": [
"GSAForCausalLM"
],
"bos_token_id": 1,
"clamp_max": null,
"clamp_min": null,
"conv_size": 4,
"elementwise_affine": false,
"eos_token_id": 2,
"expand_k": 1,
"expand_v": 1,
"feature_map": "swish",
"fuse_cross_entropy": true,
"fuse_norm": true,
"gate_logit_normalizer": 8,
"gate_low_rank_dim": null,
"hidden_act": "swish",
"hidden_ratio": 4,
"hidden_size": 2048,
"initializer_range": 0.02,
"intermediate_size": null,
"max_position_embeddings": 2048,
"model_type": "gsa",
"norm_eps": 1e-06,
"norm_first": true,
"num_heads": 4,
"num_hidden_layers": 24,
"num_kv_heads": null,
"num_slots": 64,
"pad_token_id": 2,
"share_conv_kernel": true,
"tie_word_embeddings": false,
"torch_dtype": "bfloat16",
"transformers_version": "4.40.2",
"use_cache": true,
"use_norm": true,
"use_output_gate": false,
"use_rope": false,
"use_short_conv": false,
"vocab_size": 32000
}