Upload Part T Dense seed 70 final checkpoint step 50462
Browse files- checkpoints/FINAL3_20260927/part_t/1p2b_dense_qkgainfp32_s70/checkpoint-50462-rcp/config.json +40 -0
- checkpoints/FINAL3_20260927/part_t/1p2b_dense_qkgainfp32_s70/checkpoint-50462-rcp/loader_gen_state_rank0.pt +3 -0
- checkpoints/FINAL3_20260927/part_t/1p2b_dense_qkgainfp32_s70/checkpoint-50462-rcp/loader_gen_state_rank1.pt +3 -0
- checkpoints/FINAL3_20260927/part_t/1p2b_dense_qkgainfp32_s70/checkpoint-50462-rcp/loader_gen_state_rank2.pt +3 -0
- checkpoints/FINAL3_20260927/part_t/1p2b_dense_qkgainfp32_s70/checkpoint-50462-rcp/loader_gen_state_rank3.pt +3 -0
- checkpoints/FINAL3_20260927/part_t/1p2b_dense_qkgainfp32_s70/checkpoint-50462-rcp/model.pt +3 -0
- checkpoints/FINAL3_20260927/part_t/1p2b_dense_qkgainfp32_s70/checkpoint-50462-rcp/rng_state_rank0.pt +3 -0
- checkpoints/FINAL3_20260927/part_t/1p2b_dense_qkgainfp32_s70/checkpoint-50462-rcp/rng_state_rank1.pt +3 -0
- checkpoints/FINAL3_20260927/part_t/1p2b_dense_qkgainfp32_s70/checkpoint-50462-rcp/rng_state_rank2.pt +3 -0
- checkpoints/FINAL3_20260927/part_t/1p2b_dense_qkgainfp32_s70/checkpoint-50462-rcp/rng_state_rank3.pt +3 -0
- checkpoints/FINAL3_20260927/part_t/1p2b_dense_qkgainfp32_s70/checkpoint-50462-rcp/training_state.pt +3 -0
checkpoints/FINAL3_20260927/part_t/1p2b_dense_qkgainfp32_s70/checkpoint-50462-rcp/config.json
ADDED
|
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"vocab_size": 151936,
|
| 3 |
+
"hidden_size": 1536,
|
| 4 |
+
"intermediate_size": 6144,
|
| 5 |
+
"num_hidden_layers": 28,
|
| 6 |
+
"num_attention_heads": 12,
|
| 7 |
+
"num_key_value_heads": 4,
|
| 8 |
+
"head_dim": 128,
|
| 9 |
+
"hidden_act": "silu",
|
| 10 |
+
"max_position_embeddings": 32768,
|
| 11 |
+
"initializer_range": 0.02,
|
| 12 |
+
"rms_norm_eps": 1e-06,
|
| 13 |
+
"use_cache": true,
|
| 14 |
+
"tie_word_embeddings": true,
|
| 15 |
+
"rope_theta": 1000000,
|
| 16 |
+
"rope_scaling": null,
|
| 17 |
+
"positional": "rope",
|
| 18 |
+
"qk_norm": "on",
|
| 19 |
+
"qk_gain_fp32": true,
|
| 20 |
+
"rope_train_position_scale_set": null,
|
| 21 |
+
"attention_bias": false,
|
| 22 |
+
"attention_dropout": 0.0,
|
| 23 |
+
"bos_token_id": 151643,
|
| 24 |
+
"eos_token_id": 151643,
|
| 25 |
+
"checkpoint_activations": true,
|
| 26 |
+
"attention_type": "sparse_stochastic",
|
| 27 |
+
"sparse_block_size": 128,
|
| 28 |
+
"sparse_num_random_blocks": 31,
|
| 29 |
+
"sparse_mask_pool_size": 1,
|
| 30 |
+
"sparse_max_seq_length": 8192,
|
| 31 |
+
"sparse_layer_indices": null,
|
| 32 |
+
"sparse_mask_cache_dir": null,
|
| 33 |
+
"sparse_mask_seed": 0,
|
| 34 |
+
"sparse_mixed_nrb_set": null,
|
| 35 |
+
"moba_block_size": 128,
|
| 36 |
+
"moba_top_k": 32,
|
| 37 |
+
"asentmax_alpha": 1.5,
|
| 38 |
+
"asentmax_learn_temp": true,
|
| 39 |
+
"asentmax_init_temp": 1.0
|
| 40 |
+
}
|
checkpoints/FINAL3_20260927/part_t/1p2b_dense_qkgainfp32_s70/checkpoint-50462-rcp/loader_gen_state_rank0.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d3d3d87a844d4fd6548dfc4c1235a1ee3e262506bc6260da75e7cbf9d9da819c
|
| 3 |
+
size 6802
|
checkpoints/FINAL3_20260927/part_t/1p2b_dense_qkgainfp32_s70/checkpoint-50462-rcp/loader_gen_state_rank1.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c2c029e9127ff4c346ab99e6af8377a6903003909fec0b073b70bfafc01d4517
|
| 3 |
+
size 6802
|
checkpoints/FINAL3_20260927/part_t/1p2b_dense_qkgainfp32_s70/checkpoint-50462-rcp/loader_gen_state_rank2.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:917e752ef93a10f1da3dc49f9c18adc43ae1bb568ee13d528c9824ee4513b511
|
| 3 |
+
size 6802
|
checkpoints/FINAL3_20260927/part_t/1p2b_dense_qkgainfp32_s70/checkpoint-50462-rcp/loader_gen_state_rank3.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:389dee6373df6f406c9ae897b97bc5acc1b426bc8c9810e2754f7d2f892a256f
|
| 3 |
+
size 6802
|
checkpoints/FINAL3_20260927/part_t/1p2b_dense_qkgainfp32_s70/checkpoint-50462-rcp/model.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f46af0a046305c535fe16f57aed7b16289af72a9bfe1f486f6c2178dcf304f24
|
| 3 |
+
size 2404833883
|
checkpoints/FINAL3_20260927/part_t/1p2b_dense_qkgainfp32_s70/checkpoint-50462-rcp/rng_state_rank0.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:08de481fd08cd3c1878ab188739e636191f6f866d01d31ce9ccd22538f1777e5
|
| 3 |
+
size 14821
|
checkpoints/FINAL3_20260927/part_t/1p2b_dense_qkgainfp32_s70/checkpoint-50462-rcp/rng_state_rank1.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fba7b8b43576198287654e0935c8e1ea215b501c43e9441064362eaa860e2961
|
| 3 |
+
size 14821
|
checkpoints/FINAL3_20260927/part_t/1p2b_dense_qkgainfp32_s70/checkpoint-50462-rcp/rng_state_rank2.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c34c2a2dc8506512d1a3c3edfd2f8bf9a29f85a1e677087c2a3f35e96fa09d8e
|
| 3 |
+
size 14821
|
checkpoints/FINAL3_20260927/part_t/1p2b_dense_qkgainfp32_s70/checkpoint-50462-rcp/rng_state_rank3.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:67e2783d9790e5686d5d56b9cf00fe59d116a478f7d6b025d069324c62940473
|
| 3 |
+
size 14821
|
checkpoints/FINAL3_20260927/part_t/1p2b_dense_qkgainfp32_s70/checkpoint-50462-rcp/training_state.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7ffddfd335d5b5ccfc9f53923a6d38af08f2eaa1047a51deb4cbe55e8fead629
|
| 3 |
+
size 4809712479
|