Upload Part 5 Dense 8k twin seed 42 checkpoint step 44000

checkpoints/FINAL4_20260927/part_5/365m_dense_qkgainfp32_s42_8k_twin/checkpoint-44000-tamia/config.json

@@ -0,0 +1,40 @@
+{
+  "vocab_size": 151936,
+  "hidden_size": 768,
+  "intermediate_size": 3072,
+  "num_hidden_layers": 28,
+  "num_attention_heads": 6,
+  "num_key_value_heads": 2,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "max_position_embeddings": 32768,
+  "initializer_range": 0.02,
+  "rms_norm_eps": 1e-06,
+  "use_cache": true,
+  "tie_word_embeddings": true,
+  "rope_theta": 1000000.0,
+  "rope_scaling": null,
+  "positional": "rope",
+  "qk_norm": "on",
+  "qk_gain_fp32": true,
+  "rope_train_position_scale_set": null,
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "checkpoint_activations": false,
+  "attention_type": "dense",
+  "sparse_block_size": 128,
+  "sparse_num_random_blocks": 1,
+  "sparse_mask_pool_size": 1000,
+  "sparse_max_seq_length": 8192,
+  "sparse_layer_indices": null,
+  "sparse_mask_cache_dir": null,
+  "sparse_mask_seed": 0,
+  "sparse_mixed_nrb_set": null,
+  "moba_block_size": 128,
+  "moba_top_k": 32,
+  "asentmax_alpha": 1.5,
+  "asentmax_learn_temp": true,
+  "asentmax_init_temp": 1.0
+}

checkpoints/FINAL4_20260927/part_5/365m_dense_qkgainfp32_s42_8k_twin/checkpoint-44000-tamia/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b034be702b8a3d0c65e008c3590ce966511944b1063591381e3000ee1a89587
+size 718046299