_name_or_path=/Users/jakmro/dev/CoreApp/pebble-model-eval/weights/needle-after-hf
architectures=NeedleForCausalLM
attention_head_dim=64
attention_heads=8
attention_kv_heads=4
bos_token_id=2
chunk_size_feed_forward=0
context_length=0
d_model=512
decoder_start_token_id=1
dtype=bfloat16
eos_token_id=1
ffn_intermediate_dim=0
hidden_dim=512
hidden_size=512
is_encoder_decoder=true
layer_norm_eps=1e-06
model_type=needle
moe_every_n_layers=0
num_attention_heads=8
num_decoder_layers=8
num_encoder_layers=12
num_experts=0
num_experts_per_tok=0
num_heads=8
num_hidden_layers=8
num_key_value_heads=4
num_kv_heads=4
num_layers=8
num_shared_experts=0
num_top_experts=0
output_attentions=false
output_hidden_states=false
pad_token_id=0
return_dict=true
rms_norm_eps=1e-06
rope_theta=10000.0
tie_word_embeddings=true
transformers_version=5.5.4
unk_token_id=3
vocab_size=8192
