Training in progress, step 200
Browse files
config.json
ADDED
@@ -0,0 +1,100 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"architectures": [
|
3 |
+
"HausaLMForCausalLM"
|
4 |
+
],
|
5 |
+
"model_type": "xlstm",
|
6 |
+
"text_config": {
|
7 |
+
"_block_map": "1,0,1,0,1,0,1,0,1,0,0,0,1,0,1,0,1,0,1,0",
|
8 |
+
"add_embedding_dropout": false,
|
9 |
+
"add_post_blocks_norm": true,
|
10 |
+
"bias": false,
|
11 |
+
"context_length": 128,
|
12 |
+
"dropout": 0.0,
|
13 |
+
"embedding_dim": 720,
|
14 |
+
"mlstm_block": {
|
15 |
+
"_block_idx": null,
|
16 |
+
"_num_blocks": 20,
|
17 |
+
"mlstm": {
|
18 |
+
"_inner_embedding_dim": 1472,
|
19 |
+
"_num_blocks": 20,
|
20 |
+
"_proj_up_dim": 1472,
|
21 |
+
"bias": false,
|
22 |
+
"context_length": 128,
|
23 |
+
"conv1d_kernel_size": 4,
|
24 |
+
"dropout": 0.0,
|
25 |
+
"embedding_dim": 720,
|
26 |
+
"num_heads": 4,
|
27 |
+
"proj_factor": 2.0,
|
28 |
+
"qkv_proj_blocksize": 32,
|
29 |
+
"round_proj_up_dim_up": true,
|
30 |
+
"round_proj_up_to_multiple_of": 64
|
31 |
+
}
|
32 |
+
},
|
33 |
+
"num_blocks": 20,
|
34 |
+
"slstm_at": [
|
35 |
+
0,
|
36 |
+
2,
|
37 |
+
4,
|
38 |
+
6,
|
39 |
+
8,
|
40 |
+
12,
|
41 |
+
14,
|
42 |
+
16,
|
43 |
+
18
|
44 |
+
],
|
45 |
+
"slstm_block": {
|
46 |
+
"_block_idx": null,
|
47 |
+
"_num_blocks": 20,
|
48 |
+
"feedforward": {
|
49 |
+
"_num_blocks": 1,
|
50 |
+
"_proj_up_dim": 0,
|
51 |
+
"act_fn": "swish",
|
52 |
+
"bias": false,
|
53 |
+
"dropout": 0.0,
|
54 |
+
"embedding_dim": -1,
|
55 |
+
"ff_type": "ffn_gated",
|
56 |
+
"proj_factor": 1.7,
|
57 |
+
"round_proj_up_dim_up": true,
|
58 |
+
"round_proj_up_to_multiple_of": 64
|
59 |
+
},
|
60 |
+
"slstm": {
|
61 |
+
"_block_idx": null,
|
62 |
+
"_num_blocks": 20,
|
63 |
+
"backend": "vanilla",
|
64 |
+
"batch_size": 8,
|
65 |
+
"bias_init": "powerlaw_blockdependent",
|
66 |
+
"constants": {},
|
67 |
+
"conv1d_kernel_size": 4,
|
68 |
+
"dropout": 0.0,
|
69 |
+
"dtype": "bfloat16",
|
70 |
+
"dtype_a": "float32",
|
71 |
+
"dtype_b": "float32",
|
72 |
+
"dtype_g": "bfloat16",
|
73 |
+
"dtype_r": "bfloat16",
|
74 |
+
"dtype_s": "bfloat16",
|
75 |
+
"dtype_w": "bfloat16",
|
76 |
+
"embedding_dim": 720,
|
77 |
+
"enable_automatic_mixed_precision": true,
|
78 |
+
"forward_clipval": null,
|
79 |
+
"function": "slstm",
|
80 |
+
"gradient_recurrent_clipval": null,
|
81 |
+
"gradient_recurrent_cut": false,
|
82 |
+
"group_norm_weight": true,
|
83 |
+
"hidden_size": 720,
|
84 |
+
"initial_val": 0.0,
|
85 |
+
"input_shape": "BSGNH",
|
86 |
+
"internal_input_shape": "SBNGH",
|
87 |
+
"num_gates": 4,
|
88 |
+
"num_heads": 4,
|
89 |
+
"num_states": 4,
|
90 |
+
"output_shape": "BNSH",
|
91 |
+
"recurrent_weight_init": "zeros"
|
92 |
+
}
|
93 |
+
},
|
94 |
+
"tie_weights": false,
|
95 |
+
"vocab_size": 49152,
|
96 |
+
"weight_decay_on_embedding": false
|
97 |
+
},
|
98 |
+
"torch_dtype": "float32",
|
99 |
+
"transformers_version": "4.47.0"
|
100 |
+
}
|
events.out.tfevents.1741204505.a22bb5f11002.311.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:08281589f22dacdf62920e2d2b3f90eceed7456f2f2633501f2426167382e104
|
3 |
+
size 7184
|
events.out.tfevents.1741204595.a22bb5f11002.367.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:549696cbf7dada9c0af5e3db362ab3468a3e4bbf7213a9e9566707e3ea3906ae
|
3 |
+
size 7156
|
events.out.tfevents.1741204716.a22bb5f11002.426.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fdbf26e74968b7d66fd1d90e2113c8ae40d35c4179a10cf13369e9291650fb13
|
3 |
+
size 7638
|
model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f009c73134dec2acb3482ae87b40060eb866fa803eecbd800d8925d35f305e82
|
3 |
+
size 568672088
|
tokenizer.json
CHANGED
@@ -1,7 +1,21 @@
|
|
1 |
{
|
2 |
"version": "1.0",
|
3 |
-
"truncation":
|
4 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
5 |
"added_tokens": [
|
6 |
{
|
7 |
"id": 0,
|
|
|
1 |
{
|
2 |
"version": "1.0",
|
3 |
+
"truncation": {
|
4 |
+
"direction": "Right",
|
5 |
+
"max_length": 128,
|
6 |
+
"strategy": "LongestFirst",
|
7 |
+
"stride": 0
|
8 |
+
},
|
9 |
+
"padding": {
|
10 |
+
"strategy": {
|
11 |
+
"Fixed": 128
|
12 |
+
},
|
13 |
+
"direction": "Right",
|
14 |
+
"pad_to_multiple_of": null,
|
15 |
+
"pad_id": 2,
|
16 |
+
"pad_type_id": 0,
|
17 |
+
"pad_token": "<|im_end|>"
|
18 |
+
},
|
19 |
"added_tokens": [
|
20 |
{
|
21 |
"id": 0,
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:ee6e6417d6cbd31522a69265b6233d4ea4ef190cd1ce7a59a03e9e29f7a72919
|
3 |
+
size 5688
|