Update tiny models for xLSTMForCausalLM
Browse files- config.json +43 -0
- generation_config.json +10 -0
- model.safetensors +3 -0
- tokenizer.json +0 -0
- tokenizer_config.json +15 -0
config.json
ADDED
|
@@ -0,0 +1,43 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_out_norm": true,
|
| 3 |
+
"architectures": [
|
| 4 |
+
"xLSTMForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"autocast_kernel_dtype": "bfloat16",
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"chunk_size": 64,
|
| 9 |
+
"chunkwise_kernel": "chunkwise--native_autograd",
|
| 10 |
+
"dtype": "float32",
|
| 11 |
+
"embedding_dim": 128,
|
| 12 |
+
"eos_token_id": 0,
|
| 13 |
+
"eps": 1e-06,
|
| 14 |
+
"ffn_proj_factor": 2.667,
|
| 15 |
+
"ffn_round_up_to_multiple_of": 64,
|
| 16 |
+
"gate_soft_cap": 15.0,
|
| 17 |
+
"hidden_size": 128,
|
| 18 |
+
"inference_state_dtype": "float32",
|
| 19 |
+
"is_decoder": true,
|
| 20 |
+
"max_inference_chunksize": 16384,
|
| 21 |
+
"mode": "inference",
|
| 22 |
+
"model_type": "xlstm",
|
| 23 |
+
"n_positions": 512,
|
| 24 |
+
"norm_eps": 1e-06,
|
| 25 |
+
"norm_reduction_force_float32": true,
|
| 26 |
+
"num_blocks": 2,
|
| 27 |
+
"num_heads": 2,
|
| 28 |
+
"num_hidden_layers": 2,
|
| 29 |
+
"output_logit_soft_cap": 30.0,
|
| 30 |
+
"pad_token_id": 50276,
|
| 31 |
+
"qk_dim_factor": 0.5,
|
| 32 |
+
"return_last_states": true,
|
| 33 |
+
"sequence_kernel": "native_sequence__native",
|
| 34 |
+
"step_kernel": "native",
|
| 35 |
+
"tie_word_embeddings": false,
|
| 36 |
+
"transformers_version": "5.16.0.dev0",
|
| 37 |
+
"type_vocab_size": 16,
|
| 38 |
+
"use_bias": false,
|
| 39 |
+
"use_cache": true,
|
| 40 |
+
"v_dim_factor": 1.0,
|
| 41 |
+
"vocab_size": 50277,
|
| 42 |
+
"weight_mode": "single"
|
| 43 |
+
}
|
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": 50276,
|
| 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 sha256:73220a9a5193eb8f8d87dd145a2ec2cbfce3df0d56e9fc317a7ccd1a7c3c4428
|
| 3 |
+
size 53198888
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": "<|endoftext|>",
|
| 5 |
+
"clean_up_tokenization_spaces": false,
|
| 6 |
+
"eos_token": "<|endoftext|>",
|
| 7 |
+
"errors": "replace",
|
| 8 |
+
"is_local": true,
|
| 9 |
+
"local_files_only": false,
|
| 10 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 11 |
+
"pad_token": null,
|
| 12 |
+
"tokenizer_class": "GPTNeoXTokenizer",
|
| 13 |
+
"trim_offsets": true,
|
| 14 |
+
"unk_token": "<|endoftext|>"
|
| 15 |
+
}
|