Upload 2 files
Browse files- config.json +33 -0
- zeta25b_v4_expanded_FINAL.pt +3 -0
config.json
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model_type": "rth_tcn",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"ZetaGrid25B"
|
| 5 |
+
],
|
| 6 |
+
"num_parameters": 25000000000,
|
| 7 |
+
"hidden_size": 4096,
|
| 8 |
+
"intermediate_size": 16384,
|
| 9 |
+
"num_hidden_layers": 32,
|
| 10 |
+
"vocab_size": 256,
|
| 11 |
+
"lora_rank": 128,
|
| 12 |
+
"kernel_size": 3,
|
| 13 |
+
"dilation_cycle": [
|
| 14 |
+
1,
|
| 15 |
+
2,
|
| 16 |
+
4,
|
| 17 |
+
8,
|
| 18 |
+
16,
|
| 19 |
+
32,
|
| 20 |
+
64,
|
| 21 |
+
128
|
| 22 |
+
],
|
| 23 |
+
"max_position_embeddings": 2048,
|
| 24 |
+
"torch_dtype": "bfloat16",
|
| 25 |
+
"transformers_version": "custom",
|
| 26 |
+
"model_version": "v3-knowledge",
|
| 27 |
+
"genome_params": 6400000000,
|
| 28 |
+
"trainable_params": 244000000,
|
| 29 |
+
"effective_params": 25000000000,
|
| 30 |
+
"license": "cc-by-nc-4.0",
|
| 31 |
+
"author": "Christian Quintino De Luca — RTH Italia",
|
| 32 |
+
"description": "RTH-LM 25B: Fractal Gated Causal TCN. Non-Transformer. Genome (frozen) + Soul (LoRA adapters). Linear O(N) inference."
|
| 33 |
+
}
|
zeta25b_v4_expanded_FINAL.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:416928ab45a16ead486bac3a48823791f4221105dcd411d7afb411fcb81b319d
|
| 3 |
+
size 3796449578
|