diff --git a/config/accelerate.yaml b/config/accelerate.yaml
new file mode 100644
index 0000000..066be39
--- /dev/null
+++ b/config/accelerate.yaml
@@ -0,0 +1,16 @@
+compute_environment: LOCAL_MACHINE
+debug: false
+distributed_type: 'NO'
+downcast_bf16: 'no'
+gpu_ids: all
+machine_rank: 0
+main_training_function: main
+mixed_precision: bf16
+num_machines: 1
+num_processes: 1
+rdzv_backend: static
+same_network: true
+tpu_env: []
+tpu_use_cluster: false
+tpu_use_sudo: false
+use_cpu: false
diff --git a/config/models/large-new.json b/config/models/large-new.json
deleted file mode 100644
index 44014f3..0000000
--- a/config/models/large-new.json
+++ /dev/null
@@ -1,9 +0,0 @@
-{
-    "d_model": 2048,
-    "n_heads": 32,
-    "n_layers": 16,
-    "ff_mult": 4,
-    "drop_p": 0.0,
-    "max_seq_len": 8192,
-    "grad_checkpoint": true
-}
\ No newline at end of file
diff --git a/config/models/large.json b/config/models/large.json
index 5dd03bf..44014f3 100644
--- a/config/models/large.json
+++ b/config/models/large.json
@@ -1,9 +1,9 @@
 {
-    "d_model": 1024,
-    "n_heads": 16,
-    "n_layers": 64,
+    "d_model": 2048,
+    "n_heads": 32,
+    "n_layers": 16,
     "ff_mult": 4,
     "drop_p": 0.0,
-    "max_seq_len": 4096,
+    "max_seq_len": 8192,
     "grad_checkpoint": true
 }
\ No newline at end of file
diff --git a/config/models/medium.json b/config/models/medium.json
index f765255..a1df8a6 100644
--- a/config/models/medium.json
+++ b/config/models/medium.json
@@ -1,9 +1,9 @@
 {
-    "d_model": 768,
-    "n_heads": 12,
-    "n_layers": 48,
+    "d_model": 1536,
+    "n_heads": 24,
+    "n_layers": 16,
     "ff_mult": 4,
     "drop_p": 0.0,
-    "max_seq_len": 4096,
+    "max_seq_len": 8192,
     "grad_checkpoint": true
-}
\ No newline at end of file
+}
diff --git a/config/models/small.json b/config/models/small.json
deleted file mode 100644
index 0eb6aad..0000000
--- a/config/models/small.json
+++ /dev/null
@@ -1,9 +0,0 @@
-{
-    "d_model": 512,
-    "n_heads": 8,
-    "n_layers": 32,
-    "ff_mult": 4,
-    "drop_p": 0.0,
-    "max_seq_len": 4096,
-    "grad_checkpoint": true
-}
\ No newline at end of file
diff --git a/config/models/test.json b/config/models/test.json
deleted file mode 100644
index 1099634..0000000
--- a/config/models/test.json
+++ /dev/null
@@ -1,9 +0,0 @@
-{
-    "d_model": 128,
-    "n_heads": 4,
-    "n_layers": 4,
-    "ff_mult": 4,
-    "drop_p": 0.1,
-    "max_seq_len": 256,
-    "grad_checkpoint": false
-}
\ No newline at end of file