exp010: head_apmix
Browse files
experiments/exp_010_component_potentials/runs/head_apmix/checkpoint.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:44e57d1c767d0682b77b271adc9b7263c877ad641a79c23a01d2a396667485b5
|
| 3 |
+
size 28205498
|
experiments/exp_010_component_potentials/runs/head_apmix/result.json
ADDED
|
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"run": "head_apmix",
|
| 3 |
+
"family": "control",
|
| 4 |
+
"dev_final": 0.0114976167678833,
|
| 5 |
+
"rank_final": 3.950249671936035,
|
| 6 |
+
"statute_final": "uniform",
|
| 7 |
+
"params": 6951239,
|
| 8 |
+
"minutes": 6.6,
|
| 9 |
+
"bpb": 2.9165386622613942,
|
| 10 |
+
"bpb_eval": NaN,
|
| 11 |
+
"coverage": 0.8077392578125,
|
| 12 |
+
"gate_acc": 0.8555908203125,
|
| 13 |
+
"mode_spread_deg": 86.75912555463131,
|
| 14 |
+
"mix_entropy": 0.9566619225742653,
|
| 15 |
+
"write_kl": 0.13903652131557465,
|
| 16 |
+
"kernel_aux": NaN,
|
| 17 |
+
"loss": 6.078675270080566,
|
| 18 |
+
"condition": {
|
| 19 |
+
"tag": "head_apmix",
|
| 20 |
+
"dim": "384",
|
| 21 |
+
"n_layers": "4",
|
| 22 |
+
"n_heads": "6",
|
| 23 |
+
"K": "64",
|
| 24 |
+
"seq_len": "256",
|
| 25 |
+
"batch_size": "32",
|
| 26 |
+
"accum_steps": "1",
|
| 27 |
+
"d_point": "48",
|
| 28 |
+
"head": "hybrid",
|
| 29 |
+
"steps": "5000",
|
| 30 |
+
"log_every": "500",
|
| 31 |
+
"bank_scorer": "apmix",
|
| 32 |
+
"n_pointers": "4",
|
| 33 |
+
"bank_source": "atlas/bank_16384_tri.pt"
|
| 34 |
+
}
|
| 35 |
+
}
|
experiments/exp_010_component_potentials/runs/head_apmix/snapshots.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b3b077d1506f40a39738f2b2c7561f51a555486b31d8fe8cb916134a9479e668
|
| 3 |
+
size 17188
|