exp010: ctx_16x256
Browse files
experiments/exp_010_component_potentials/runs/ctx_16x256/checkpoint.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f38789f7870e42bf127531e34691b3f43f6ad5a0220e804f94117a93bc39b064
|
| 3 |
+
size 27303790
|
experiments/exp_010_component_potentials/runs/ctx_16x256/result.json
ADDED
|
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"run": "ctx_16x256",
|
| 3 |
+
"family": "ctx",
|
| 4 |
+
"dev_final": 0.008060336112976074,
|
| 5 |
+
"rank_final": 3.896486282348633,
|
| 6 |
+
"statute_final": "uniform",
|
| 7 |
+
"params": 6753090,
|
| 8 |
+
"minutes": 87.0,
|
| 9 |
+
"bpb": 3.0981225312463705,
|
| 10 |
+
"bpb_eval": NaN,
|
| 11 |
+
"coverage": 0.931640625,
|
| 12 |
+
"gate_acc": 0.93408203125,
|
| 13 |
+
"mode_spread_deg": NaN,
|
| 14 |
+
"mix_entropy": NaN,
|
| 15 |
+
"write_kl": 0.14831112325191498,
|
| 16 |
+
"kernel_aux": NaN,
|
| 17 |
+
"loss": 6.306842356920242,
|
| 18 |
+
"condition": {
|
| 19 |
+
"tag": "ctx_16x256",
|
| 20 |
+
"dim": "384",
|
| 21 |
+
"n_layers": "4",
|
| 22 |
+
"n_heads": "6",
|
| 23 |
+
"K": "64",
|
| 24 |
+
"seq_len": "256",
|
| 25 |
+
"batch_size": "8",
|
| 26 |
+
"accum_steps": "1",
|
| 27 |
+
"d_point": "48",
|
| 28 |
+
"head": "hybrid",
|
| 29 |
+
"steps": "5000",
|
| 30 |
+
"log_every": "500",
|
| 31 |
+
"train_mode": "stream",
|
| 32 |
+
"pos_mode": "clamp",
|
| 33 |
+
"segments": "16"
|
| 34 |
+
}
|
| 35 |
+
}
|
experiments/exp_010_component_potentials/runs/ctx_16x256/snapshots.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3ab1f69305f94ebdba8923f3370fced6442a0defcb99177b72b396733ea745fc
|
| 3 |
+
size 17188
|