Update tiny models for MellumForCausalLM
Browse files- config.json +52 -0
- generation_config.json +10 -0
- model.safetensors +3 -0
- tokenizer.json +0 -0
- tokenizer_config.json +12 -0
config.json
ADDED
|
@@ -0,0 +1,52 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"MellumForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_bias": false,
|
| 6 |
+
"attention_dropout": 0.0,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"dtype": "float32",
|
| 9 |
+
"eos_token_id": 0,
|
| 10 |
+
"head_dim": 16,
|
| 11 |
+
"hidden_act": "gelu",
|
| 12 |
+
"hidden_size": 32,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 32,
|
| 15 |
+
"is_decoder": true,
|
| 16 |
+
"layer_types": [
|
| 17 |
+
"full_attention",
|
| 18 |
+
"sliding_attention"
|
| 19 |
+
],
|
| 20 |
+
"max_position_embeddings": 512,
|
| 21 |
+
"mlp_layer_types": [
|
| 22 |
+
"dense",
|
| 23 |
+
"sparse"
|
| 24 |
+
],
|
| 25 |
+
"model_type": "mellum",
|
| 26 |
+
"moe_intermediate_size": 16,
|
| 27 |
+
"norm_topk_prob": true,
|
| 28 |
+
"num_attention_heads": 2,
|
| 29 |
+
"num_experts_per_tok": 2,
|
| 30 |
+
"num_hidden_layers": 2,
|
| 31 |
+
"num_key_value_heads": 2,
|
| 32 |
+
"num_local_experts": 8,
|
| 33 |
+
"output_router_logits": false,
|
| 34 |
+
"pad_token_id": 0,
|
| 35 |
+
"rms_norm_eps": 1e-06,
|
| 36 |
+
"rope_parameters": {
|
| 37 |
+
"full_attention": {
|
| 38 |
+
"rope_theta": 500000.0,
|
| 39 |
+
"rope_type": "default"
|
| 40 |
+
},
|
| 41 |
+
"sliding_attention": {
|
| 42 |
+
"rope_theta": 10000.0,
|
| 43 |
+
"rope_type": "default"
|
| 44 |
+
}
|
| 45 |
+
},
|
| 46 |
+
"router_aux_loss_coef": 0.001,
|
| 47 |
+
"sliding_window": 1024,
|
| 48 |
+
"tie_word_embeddings": false,
|
| 49 |
+
"transformers_version": "5.16.0.dev0",
|
| 50 |
+
"use_cache": true,
|
| 51 |
+
"vocab_size": 98304
|
| 52 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 0,
|
| 4 |
+
"eos_token_id": 0,
|
| 5 |
+
"output_attentions": false,
|
| 6 |
+
"output_hidden_states": false,
|
| 7 |
+
"pad_token_id": 0,
|
| 8 |
+
"transformers_version": "5.16.0.dev0",
|
| 9 |
+
"use_cache": true
|
| 10 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9d126201a070b48fa4744de09c5ddde82dc916a5d9d3090463defee18865905c
|
| 3 |
+
size 25267128
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"backend": "tokenizers",
|
| 3 |
+
"bos_token": "<|endoftext|>",
|
| 4 |
+
"clean_up_tokenization_spaces": false,
|
| 5 |
+
"eos_token": "<|endoftext|>",
|
| 6 |
+
"is_local": true,
|
| 7 |
+
"local_files_only": false,
|
| 8 |
+
"model_max_length": 512,
|
| 9 |
+
"pad_token": "<|endoftext|>",
|
| 10 |
+
"tokenizer_class": "TokenizersBackend",
|
| 11 |
+
"unk_token": "<|endoftext|>"
|
| 12 |
+
}
|