Training in progress, step 500
Browse files
config.json
ADDED
|
@@ -0,0 +1,70 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"Llama4ForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_bias": false,
|
| 6 |
+
"attention_chunk_size": 8192,
|
| 7 |
+
"attention_dropout": 0.0,
|
| 8 |
+
"attn_scale": 0.1,
|
| 9 |
+
"attn_temperature_tuning": true,
|
| 10 |
+
"bos_token_id": 4,
|
| 11 |
+
"eos_token_id": 5,
|
| 12 |
+
"floor_scale": 8192,
|
| 13 |
+
"head_dim": 128,
|
| 14 |
+
"hidden_act": "silu",
|
| 15 |
+
"hidden_size": 512,
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"interleave_moe_layer_step": 1,
|
| 18 |
+
"intermediate_size": 2048,
|
| 19 |
+
"intermediate_size_mlp": 16384,
|
| 20 |
+
"layer_types": [
|
| 21 |
+
"chunked_attention",
|
| 22 |
+
"chunked_attention",
|
| 23 |
+
"chunked_attention",
|
| 24 |
+
"full_attention",
|
| 25 |
+
"chunked_attention",
|
| 26 |
+
"chunked_attention",
|
| 27 |
+
"chunked_attention",
|
| 28 |
+
"full_attention"
|
| 29 |
+
],
|
| 30 |
+
"max_position_embeddings": 2048,
|
| 31 |
+
"model_type": "llama4_text",
|
| 32 |
+
"moe_layers": [
|
| 33 |
+
0,
|
| 34 |
+
1,
|
| 35 |
+
2,
|
| 36 |
+
3,
|
| 37 |
+
4,
|
| 38 |
+
5,
|
| 39 |
+
6,
|
| 40 |
+
7
|
| 41 |
+
],
|
| 42 |
+
"no_rope_layers": [
|
| 43 |
+
1,
|
| 44 |
+
1,
|
| 45 |
+
1,
|
| 46 |
+
0,
|
| 47 |
+
1,
|
| 48 |
+
1,
|
| 49 |
+
1,
|
| 50 |
+
0
|
| 51 |
+
],
|
| 52 |
+
"num_attention_heads": 8,
|
| 53 |
+
"num_experts_per_tok": 1,
|
| 54 |
+
"num_hidden_layers": 8,
|
| 55 |
+
"num_key_value_heads": 8,
|
| 56 |
+
"num_local_experts": 16,
|
| 57 |
+
"output_router_logits": false,
|
| 58 |
+
"pad_token_id": 0,
|
| 59 |
+
"rms_norm_eps": 1e-05,
|
| 60 |
+
"rope_scaling": null,
|
| 61 |
+
"rope_theta": 500000,
|
| 62 |
+
"router_aux_loss_coef": 0.001,
|
| 63 |
+
"router_jitter_noise": 0.0,
|
| 64 |
+
"tie_word_embeddings": false,
|
| 65 |
+
"torch_dtype": "float32",
|
| 66 |
+
"transformers_version": "4.53.2",
|
| 67 |
+
"use_cache": true,
|
| 68 |
+
"use_qk_norm": true,
|
| 69 |
+
"vocab_size": 1065
|
| 70 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d34244aa23affa7866cc41ec6c17c21dae049f6d456609362ac440ceda749613
|
| 3 |
+
size 1783055976
|
runs/Jul17_15-02-42_f523c4bc5fb7/events.out.tfevents.1752764564.f523c4bc5fb7.9633.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:beee5ae0f9f1cb607bfa99d3f39a30e4da2dcafed8db11fae2116bbf94fb37bf
|
| 3 |
+
size 5771
|
runs/Jul17_15-36-57_f523c4bc5fb7/events.out.tfevents.1752766619.f523c4bc5fb7.19217.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:72a9c45d4e209b99c431e2f353c961634d654119a42c8b3864f8349d2539e6cc
|
| 3 |
+
size 8144
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:65986360466587c8fea248a9270c0a0a84e47ac9bbbbc82defb00c4a4dc06b45
|
| 3 |
+
size 5432
|