inference/research
verifiedhigh · 0.90updated 2026-10-04

GLM-4.7-Flash config.json

repo

Fields

id
sources/mbl-cfg-glm-4-7-flash
Publisher
Hugging Face
Kind
repo
Accessed
2026-10-04
notes
glm4_moe_lite, Glm4MoeLiteForCausalLM. num_hidden_layers 47, hidden_size 2048, num_attention_heads 20, num_key_value_heads 20, kv_lora_rank 512, q_lora_rank 768, qk_nope_head_dim 192, qk_rope_head_dim 64, v_head_dim 256, partial_rotary_factor 1.0. MoE: n_routed_experts 64, n_shared_experts 1, num_experts_per_tok 4, moe_intermediate_size 1536, first_k_dense_replace 1. rope_scaling null, rope_theta 1000000. max_position_embeddings 202752. Read raw 2026-10-04.

Notes

glm4_moe_lite, Glm4MoeLiteForCausalLM. num_hidden_layers 47, hidden_size 2048, num_attention_heads 20, num_key_value_heads 20, kv_lora_rank 512, q_lora_rank 768, qk_nope_head_dim 192, qk_rope_head_dim 64, v_head_dim 256, partial_rotary_factor 1.0. MoE: n_routed_experts 64, n_shared_experts 1, num_experts_per_tok 4, moe_intermediate_size 1536, first_k_dense_replace 1. rope_scaling null, rope_theta 1000000. max_position_embeddings 202752. Read raw 2026-10-04.

Cross-references

This record references no other records.