Upload folder using huggingface_hub

snake-1584f975-checkpoint-3/config.h

@@ -0,0 +1,9 @@
+#pragma once
+
+#define N_LAYERS 12
+#define VOCAB_SIZE 256
+#define D_MODEL 768
+#define D_INNER 1536
+#define DT_RANK 48
+#define D_STATE 16
+#define D_CONV 4

snake-1584f975-checkpoint-3/config.json

@@ -0,0 +1 @@
+{"d_model": 768, "n_layer": 12, "vocab_size": 256, "ssm_cfg": {}, "rms_norm": true, "residual_in_fp32": true, "fused_add_norm": true, "pad_vocab_size_multiple": 8, "tie_embeddings": false}

snake-1584f975-checkpoint-3/model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e269ebb5be9249c6d3ae38ed2ab3405db4ad66bbbbb8ac978f61defb19af0ae6
+size 182615040

snake-1584f975-checkpoint-3/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b40847f4b96cb279e2d4782f51377a06c6905ed921bc4c2385ae7bba7307af1e
+size 182661686

snake-1584f975-checkpoint-3/training_params.json

@@ -0,0 +1,18 @@
+{
+    "d_model": 768,
+    "n_layers": 12,
+    "vocab_size": 256,
+    "model_size": 45653760,
+    "learning_rate": 1e-05,
+    "seq_length": 1024,
+    "batch_size": 4,
+    "num_batches": 1241,
+    "tokenizer_path": "",
+    "dataset_path": "flpelerin/TinyStories-train-1M",
+    "model_name": "snake-1584f975",
+    "wandb_project": "snake-training",
+    "wandb_user": "florianpelerin110304",
+    "hf_repo": "flpelerin/snakes",
+    "target_epoch": 4,
+    "current_epoch": 3
+}