DBTrimKV-Qwen3-4B-Math / config.json
ngocbh's picture
Upload DBTrimKV checkpoint (Qwen3-4B, OpenR1-Math-220k)
fe37a5f verified
{
"alpha_threshold": 0.0,
"architectures": [
"TrimKVQwen3ForCausalLM"
],
"attention_bias": false,
"attention_dropout": 0.0,
"attn_impl": "rg_attn_flex",
"base_loss": "fwkl_ntp",
"base_model": "Qwen/Qwen3-4B",
"bos_token_id": 151643,
"buffer_size": 128,
"compress_memory": true,
"compress_strategy": "alpha",
"dtype": "bfloat16",
"eos_token_id": 151645,
"floor_budget_ratio": 0.0,
"global_capacity": true,
"head_dim": 128,
"hidden_act": "silu",
"hidden_size": 2560,
"initializer_range": 0.02,
"intermediate_size": 9728,
"layer_types": [
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention",
"full_attention"
],
"logit_block_size": 16384,
"max_position_embeddings": 40960,
"max_seq_len": 32768,
"max_window_layers": 36,
"memory_size": 128.0,
"model_type": "qwen3",
"num_attention_heads": 32,
"num_hidden_layers": 36,
"num_key_value_heads": 8,
"retention_gate": "rg10",
"retention_gate_bias_init": 18.0,
"retention_gate_intermediate_size": 512,
"retention_weight": 1.0,
"rg_dropout": 0.0,
"rms_norm_eps": 1e-06,
"rope_scaling": null,
"rope_theta": 1000000,
"sliding_window": null,
"tie_retention_gate_layers": true,
"tie_word_embeddings": true,
"trainable_params": "self_attn.retention_gate",
"transformers_version": "4.57.1",
"use_cache": false,
"use_sliding_window": false,
"vocab_size": 151936
}