Training in progress, step 300

config.json

@@ -15,7 +15,7 @@
   "n_embd": 768,
   "n_head": 3,
   "n_inner": null,
-  "n_layer": 1,
+  "n_layer": 3,
   "n_positions": 1024,
   "reorder_and_upcast_attn": false,
   "resid_pdrop": 0.1,

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dfcd21b985f54b7f6f0d52f2b470a43056d6b28b968459cbc1fae1b44cbf9b33
-size 31554088
+oid sha256:b1f6af0d54ca60c27f899987ac3f3fc0815faa78bcebfedb8e7579a269cded32
+size 88259456

runs/Mar17_17-38-08_DESKTOP-C4VDTPF/events.out.tfevents.1710686290.DESKTOP-C4VDTPF.28456.8

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0fee0ab2e99d1f83a68aff2ee535f5d557f048c795a8b8a054219f88ca232938
-size 26408
+oid sha256:22c2b0e3896d83a80437eb0d64e076282417372a247d11860dce07cda304ab55
+size 27426

runs/Mar17_19-03-10_DESKTOP-C4VDTPF/events.out.tfevents.1710691391.DESKTOP-C4VDTPF.28456.9

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3659cdf0c3ef616d3fb9319e505496ed3e69ec9b8913dc0db24edda0f13a0865
+size 4420

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5b15bef375d14a837adf354e1a80faea4902fe630e23710e9654c01cd027213b
+oid sha256:42d23f673639590ac339c02051a2d834b7a2197f3943642cedbd73abc5ea24e5
 size 4536