Upload folder using huggingface_hub
Browse files- config.json +45 -4
- generation_config.json +3 -5
- model.safetensors.index.json +0 -0
- tokenizer.json +4 -63
config.json
CHANGED
@@ -1,10 +1,51 @@
|
|
1 |
{
|
|
|
2 |
"architectures": [
|
3 |
"LlamaForCausalLM"
|
4 |
],
|
5 |
"attention_bias": false,
|
6 |
"attention_dropout": 0.0,
|
7 |
"bos_token_id": 128000,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
8 |
"eos_token_id": [
|
9 |
128001,
|
10 |
128008,
|
@@ -24,15 +65,15 @@
|
|
24 |
"rms_norm_eps": 1e-05,
|
25 |
"rope_scaling": {
|
26 |
"factor": 8.0,
|
27 |
-
"low_freq_factor": 1.0,
|
28 |
"high_freq_factor": 4.0,
|
|
|
29 |
"original_max_position_embeddings": 8192,
|
30 |
"rope_type": "llama3"
|
31 |
},
|
32 |
"rope_theta": 500000.0,
|
33 |
"tie_word_embeddings": false,
|
34 |
-
"torch_dtype": "
|
35 |
-
"transformers_version": "4.
|
36 |
"use_cache": true,
|
37 |
"vocab_size": 128256
|
38 |
-
}
|
|
|
1 |
{
|
2 |
+
"_name_or_path": "/home/meta-llama/Meta-Llama-3.1-405B-Instruct",
|
3 |
"architectures": [
|
4 |
"LlamaForCausalLM"
|
5 |
],
|
6 |
"attention_bias": false,
|
7 |
"attention_dropout": 0.0,
|
8 |
"bos_token_id": 128000,
|
9 |
+
"compression_config": {
|
10 |
+
"config_groups": {
|
11 |
+
"group_0": {
|
12 |
+
"input_activations": {
|
13 |
+
"block_structure": null,
|
14 |
+
"dynamic": true,
|
15 |
+
"group_size": null,
|
16 |
+
"num_bits": 8,
|
17 |
+
"observer": "memoryless",
|
18 |
+
"observer_kwargs": {},
|
19 |
+
"strategy": "token",
|
20 |
+
"symmetric": true,
|
21 |
+
"type": "float"
|
22 |
+
},
|
23 |
+
"output_activations": null,
|
24 |
+
"targets": [
|
25 |
+
"Linear"
|
26 |
+
],
|
27 |
+
"weights": {
|
28 |
+
"block_structure": null,
|
29 |
+
"dynamic": false,
|
30 |
+
"group_size": null,
|
31 |
+
"num_bits": 8,
|
32 |
+
"observer": "minmax",
|
33 |
+
"observer_kwargs": {},
|
34 |
+
"strategy": "channel",
|
35 |
+
"symmetric": true,
|
36 |
+
"type": "float"
|
37 |
+
}
|
38 |
+
}
|
39 |
+
},
|
40 |
+
"format": "naive-quantized",
|
41 |
+
"global_compression_ratio": 1.240844659615461,
|
42 |
+
"ignore": [
|
43 |
+
"lm_head"
|
44 |
+
],
|
45 |
+
"kv_cache_scheme": null,
|
46 |
+
"quant_method": "compressed-tensors",
|
47 |
+
"quantization_status": "frozen"
|
48 |
+
},
|
49 |
"eos_token_id": [
|
50 |
128001,
|
51 |
128008,
|
|
|
65 |
"rms_norm_eps": 1e-05,
|
66 |
"rope_scaling": {
|
67 |
"factor": 8.0,
|
|
|
68 |
"high_freq_factor": 4.0,
|
69 |
+
"low_freq_factor": 1.0,
|
70 |
"original_max_position_embeddings": 8192,
|
71 |
"rope_type": "llama3"
|
72 |
},
|
73 |
"rope_theta": 500000.0,
|
74 |
"tie_word_embeddings": false,
|
75 |
+
"torch_dtype": "float16",
|
76 |
+
"transformers_version": "4.43.1",
|
77 |
"use_cache": true,
|
78 |
"vocab_size": 128256
|
79 |
+
}
|
generation_config.json
CHANGED
@@ -1,12 +1,10 @@
|
|
1 |
{
|
|
|
2 |
"bos_token_id": 128000,
|
3 |
-
"do_sample": true,
|
4 |
"eos_token_id": [
|
5 |
128001,
|
6 |
128008,
|
7 |
128009
|
8 |
],
|
9 |
-
"
|
10 |
-
|
11 |
-
"transformers_version": "4.42.3"
|
12 |
-
}
|
|
|
1 |
{
|
2 |
+
"_from_model_config": true,
|
3 |
"bos_token_id": 128000,
|
|
|
4 |
"eos_token_id": [
|
5 |
128001,
|
6 |
128008,
|
7 |
128009
|
8 |
],
|
9 |
+
"transformers_version": "4.43.1"
|
10 |
+
}
|
|
|
|
model.safetensors.index.json
CHANGED
The diff for this file is too large to render.
See raw diff
|
|
tokenizer.json
CHANGED
@@ -2329,69 +2329,10 @@
|
|
2329 |
]
|
2330 |
},
|
2331 |
"post_processor": {
|
2332 |
-
"type": "
|
2333 |
-
"
|
2334 |
-
|
2335 |
-
|
2336 |
-
"add_prefix_space": true,
|
2337 |
-
"trim_offsets": false,
|
2338 |
-
"use_regex": true
|
2339 |
-
},
|
2340 |
-
{
|
2341 |
-
"type": "TemplateProcessing",
|
2342 |
-
"single": [
|
2343 |
-
{
|
2344 |
-
"SpecialToken": {
|
2345 |
-
"id": "<|begin_of_text|>",
|
2346 |
-
"type_id": 0
|
2347 |
-
}
|
2348 |
-
},
|
2349 |
-
{
|
2350 |
-
"Sequence": {
|
2351 |
-
"id": "A",
|
2352 |
-
"type_id": 0
|
2353 |
-
}
|
2354 |
-
}
|
2355 |
-
],
|
2356 |
-
"pair": [
|
2357 |
-
{
|
2358 |
-
"SpecialToken": {
|
2359 |
-
"id": "<|begin_of_text|>",
|
2360 |
-
"type_id": 0
|
2361 |
-
}
|
2362 |
-
},
|
2363 |
-
{
|
2364 |
-
"Sequence": {
|
2365 |
-
"id": "A",
|
2366 |
-
"type_id": 0
|
2367 |
-
}
|
2368 |
-
},
|
2369 |
-
{
|
2370 |
-
"SpecialToken": {
|
2371 |
-
"id": "<|begin_of_text|>",
|
2372 |
-
"type_id": 1
|
2373 |
-
}
|
2374 |
-
},
|
2375 |
-
{
|
2376 |
-
"Sequence": {
|
2377 |
-
"id": "B",
|
2378 |
-
"type_id": 1
|
2379 |
-
}
|
2380 |
-
}
|
2381 |
-
],
|
2382 |
-
"special_tokens": {
|
2383 |
-
"<|begin_of_text|>": {
|
2384 |
-
"id": "<|begin_of_text|>",
|
2385 |
-
"ids": [
|
2386 |
-
128000
|
2387 |
-
],
|
2388 |
-
"tokens": [
|
2389 |
-
"<|begin_of_text|>"
|
2390 |
-
]
|
2391 |
-
}
|
2392 |
-
}
|
2393 |
-
}
|
2394 |
-
]
|
2395 |
},
|
2396 |
"decoder": {
|
2397 |
"type": "ByteLevel",
|
|
|
2329 |
]
|
2330 |
},
|
2331 |
"post_processor": {
|
2332 |
+
"type": "ByteLevel",
|
2333 |
+
"add_prefix_space": true,
|
2334 |
+
"trim_offsets": false,
|
2335 |
+
"use_regex": true
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2336 |
},
|
2337 |
"decoder": {
|
2338 |
"type": "ByteLevel",
|