Hugging Face's logo

alphatozeta
/
tinyllama_no_tokenizer

Model card Files
xet
Community
alphatozeta commited on
Commit
d2de6f1
·
verified ·
1 Parent(s): 727b83a

Upload folder using huggingface_hub

Browse files
Files changed (4) hide show
  1. .gitattributes +1 -0
  2. briton_config.pbtxt +0 -0
  3. config.json +231 -0
  4. rank0.engine +3 -0
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ rank0.engine filter=lfs diff=lfs merge=lfs -text
briton_config.pbtxt ADDED
The diff for this file is too large to render. See raw diff
 
config.json ADDED
@@ -0,0 +1,231 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "version": "0.17.0.post1",
3
+ "pretrained_config": {
4
+ "mlp_bias": false,
5
+ "attn_bias": false,
6
+ "rotary_base": 10000.0,
7
+ "rotary_scaling": null,
8
+ "residual_mlp": false,
9
+ "disable_weight_only_quant_plugin": false,
10
+ "moe": {
11
+ "num_experts": 0,
12
+ "shared_expert_intermediate_size": 0,
13
+ "top_k": 0,
14
+ "normalization_mode": null,
15
+ "sparse_mixer_epsilon": 0.01,
16
+ "tp_mode": 0,
17
+ "device_limited_n_group": 0,
18
+ "device_limited_topk_group": 0,
19
+ "device_limited_routed_scaling_factor": 1.0
20
+ },
21
+ "remove_duplicated_kv_heads": false,
22
+ "fc_after_embed": false,
23
+ "use_input_layernorm_in_first_layer": true,
24
+ "use_last_layernorm": true,
25
+ "layer_idx_offset": 0,
26
+ "embedding_multiplier": 1.0,
27
+ "attention_multiplier": 1.0,
28
+ "residual_multiplier": 1.0,
29
+ "output_multiplier_scale": 1.0,
30
+ "architecture": "LlamaForCausalLM",
31
+ "dtype": "bfloat16",
32
+ "vocab_size": 32000,
33
+ "hidden_size": 2048,
34
+ "num_hidden_layers": 22,
35
+ "num_attention_heads": 32,
36
+ "hidden_act": "silu",
37
+ "logits_dtype": "float16",
38
+ "norm_epsilon": 1e-05,
39
+ "runtime_defaults": null,
40
+ "position_embedding_type": "rope_gpt_neox",
41
+ "num_key_value_heads": 4,
42
+ "intermediate_size": 5632,
43
+ "max_position_embeddings": 2048,
44
+ "mapping": {
45
+ "world_size": 1,
46
+ "gpus_per_node": 8,
47
+ "cp_size": 1,
48
+ "tp_size": 1,
49
+ "pp_size": 1,
50
+ "moe_tp_size": 1,
51
+ "moe_ep_size": 1,
52
+ "auto_parallel": false
53
+ },
54
+ "quantization": {
55
+ "quant_algo": "FP8",
56
+ "kv_cache_quant_algo": "FP8",
57
+ "group_size": 128,
58
+ "smoothquant_val": 0.5,
59
+ "clamp_val": null,
60
+ "use_meta_recipe": false,
61
+ "has_zero_point": false,
62
+ "pre_quant_scale": false,
63
+ "exclude_modules": [
64
+ "transformer.layers.14.post_layernorm",
65
+ "transformer.layers.4.post_layernorm",
66
+ "transformer.layers.2.input_layernorm",
67
+ "transformer.vocab_embedding",
68
+ "transformer.layers.7.input_layernorm",
69
+ "transformer.layers.18.input_layernorm",
70
+ "lm_head",
71
+ "transformer.layers.20.input_layernorm",
72
+ "transformer.layers.10.post_layernorm",
73
+ "transformer.layers.16.input_layernorm",
74
+ "transformer.layers.8.input_layernorm",
75
+ "transformer.layers.1.input_layernorm",
76
+ "transformer.layers.20.post_layernorm",
77
+ "transformer.layers.14.input_layernorm",
78
+ "transformer.layers.3.post_layernorm",
79
+ "transformer.layers.17.input_layernorm",
80
+ "transformer.layers.17.post_layernorm",
81
+ "transformer.layers.13.input_layernorm",
82
+ "transformer.layers.21.input_layernorm",
83
+ "transformer.layers.5.input_layernorm",
84
+ "transformer.layers.6.post_layernorm",
85
+ "transformer.layers.19.input_layernorm",
86
+ "transformer.layers.12.post_layernorm",
87
+ "transformer.layers.21.post_layernorm",
88
+ "transformer.layers.11.input_layernorm",
89
+ "transformer.layers.0.post_layernorm",
90
+ "transformer.layers.16.post_layernorm",
91
+ "transformer.layers.4.input_layernorm",
92
+ "transformer.layers.11.post_layernorm",
93
+ "transformer.layers.8.post_layernorm",
94
+ "transformer.layers.6.input_layernorm",
95
+ "transformer.layers.3.input_layernorm",
96
+ "transformer.layers.10.input_layernorm",
97
+ "transformer.layers.1.post_layernorm",
98
+ "transformer.layers.15.input_layernorm",
99
+ "transformer.layers.0.input_layernorm",
100
+ "transformer.layers.9.input_layernorm",
101
+ "transformer.layers.15.post_layernorm",
102
+ "transformer.layers.18.post_layernorm",
103
+ "transformer.layers.2.post_layernorm",
104
+ "transformer.layers.13.post_layernorm",
105
+ "transformer.layers.7.post_layernorm",
106
+ "transformer.layers.12.input_layernorm",
107
+ "transformer.ln_f",
108
+ "transformer.layers.9.post_layernorm",
109
+ "transformer.layers.5.post_layernorm",
110
+ "transformer.layers.19.post_layernorm"
111
+ ]
112
+ },
113
+ "use_parallel_embedding": true,
114
+ "embedding_sharding_dim": 0,
115
+ "head_size": 64,
116
+ "qk_layernorm": false,
117
+ "rotary_embedding_dim": 64,
118
+ "producer": {
119
+ "name": "modelopt",
120
+ "version": "0.23.2"
121
+ },
122
+ "share_embedding_table": false,
123
+ "bias": false,
124
+ "rotary_pct": 1.0,
125
+ "rank": 0,
126
+ "decoder": "llama",
127
+ "rmsnorm": true,
128
+ "lm_head_bias": false,
129
+ "tie_word_embeddings": false,
130
+ "model_type": "llama"
131
+ },
132
+ "build_config": {
133
+ "max_input_len": 4096,
134
+ "max_seq_len": 4096,
135
+ "opt_batch_size": 8,
136
+ "max_batch_size": 1,
137
+ "max_beam_width": 1,
138
+ "max_num_tokens": 4096,
139
+ "opt_num_tokens": null,
140
+ "max_prompt_embedding_table_size": 0,
141
+ "kv_cache_type": "PAGED",
142
+ "gather_context_logits": false,
143
+ "gather_generation_logits": false,
144
+ "strongly_typed": true,
145
+ "force_num_profiles": null,
146
+ "profiling_verbosity": "layer_names_only",
147
+ "enable_debug_output": false,
148
+ "max_draft_len": 0,
149
+ "speculative_decoding_mode": 1,
150
+ "use_refit": false,
151
+ "input_timing_cache": null,
152
+ "output_timing_cache": "model.cache",
153
+ "lora_config": {
154
+ "lora_dir": [],
155
+ "lora_ckpt_source": "hf",
156
+ "max_lora_rank": 64,
157
+ "lora_target_modules": [],
158
+ "trtllm_modules_to_hf_modules": {}
159
+ },
160
+ "auto_parallel_config": {
161
+ "world_size": 1,
162
+ "gpus_per_node": 8,
163
+ "cluster_key": null,
164
+ "cluster_info": null,
165
+ "sharding_cost_model": "alpha_beta",
166
+ "comm_cost_model": "alpha_beta",
167
+ "enable_pipeline_parallelism": false,
168
+ "enable_shard_unbalanced_shape": false,
169
+ "enable_shard_dynamic_shape": false,
170
+ "enable_reduce_scatter": true,
171
+ "builder_flags": null,
172
+ "debug_mode": false,
173
+ "infer_shape": true,
174
+ "validation_mode": false,
175
+ "same_buffer_io": {},
176
+ "same_spec_io": {},
177
+ "sharded_io_allowlist": [],
178
+ "fill_weights": false,
179
+ "parallel_config_cache": null,
180
+ "profile_cache": null,
181
+ "dump_path": null,
182
+ "debug_outputs": []
183
+ },
184
+ "weight_sparsity": false,
185
+ "weight_streaming": false,
186
+ "plugin_config": {
187
+ "dtype": "bfloat16",
188
+ "bert_attention_plugin": "auto",
189
+ "gpt_attention_plugin": "auto",
190
+ "gemm_plugin": null,
191
+ "explicitly_disable_gemm_plugin": false,
192
+ "gemm_swiglu_plugin": null,
193
+ "fp8_rowwise_gemm_plugin": null,
194
+ "qserve_gemm_plugin": null,
195
+ "identity_plugin": null,
196
+ "nccl_plugin": null,
197
+ "lora_plugin": null,
198
+ "weight_only_groupwise_quant_matmul_plugin": null,
199
+ "weight_only_quant_matmul_plugin": null,
200
+ "smooth_quant_plugins": true,
201
+ "smooth_quant_gemm_plugin": null,
202
+ "layernorm_quantization_plugin": null,
203
+ "rmsnorm_quantization_plugin": null,
204
+ "quantize_per_token_plugin": false,
205
+ "quantize_tensor_plugin": false,
206
+ "moe_plugin": "auto",
207
+ "mamba_conv1d_plugin": "auto",
208
+ "low_latency_gemm_plugin": null,
209
+ "low_latency_gemm_swiglu_plugin": null,
210
+ "context_fmha": true,
211
+ "bert_context_fmha_fp32_acc": false,
212
+ "paged_kv_cache": true,
213
+ "remove_input_padding": true,
214
+ "reduce_fusion": false,
215
+ "user_buffer": false,
216
+ "tokens_per_block": 64,
217
+ "use_paged_context_fmha": true,
218
+ "use_fp8_context_fmha": true,
219
+ "multiple_profiles": true,
220
+ "paged_state": false,
221
+ "streamingllm": false,
222
+ "manage_weights": false,
223
+ "use_fused_mlp": true,
224
+ "pp_reduce_scatter": false
225
+ },
226
+ "use_strip_plan": false,
227
+ "max_encoder_input_len": 1024,
228
+ "monitor_memory": false,
229
+ "use_mrope": false
230
+ }
231
+ }
rank0.engine ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3df5742005247ad4cf6dff7088e9086fe3b8fb367405f223c89d1ddea19502c2
3
+ size 1280187708