Hugging Face's logo

heegyu
/
EXAONE-3.5-2.4B-Instruct-TRT

Model card Files
xet
Community
heegyu commited on
Commit
f54ee96
·
verified ·
1 Parent(s): 3b0945f

Upload folder using huggingface_hub

Browse files
Files changed (7) hide show
  1. .gitattributes +3 -0
  2. bf16/1-gpu/config.json +187 -0
  3. bf16/1-gpu/rank0.engine +3 -0
  4. int4_wq/1-gpu/config.json +187 -0
  5. int4_wq/1-gpu/rank0.engine +3 -0
  6. int8_wq/1-gpu/config.json +187 -0
  7. int8_wq/1-gpu/rank0.engine +3 -0
.gitattributes CHANGED
@@ -33,3 +33,6 @@ 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
+ bf16/1-gpu/rank0.engine filter=lfs diff=lfs merge=lfs -text
37
+ int4_wq/1-gpu/rank0.engine filter=lfs diff=lfs merge=lfs -text
38
+ int8_wq/1-gpu/rank0.engine filter=lfs diff=lfs merge=lfs -text
bf16/1-gpu/config.json ADDED
@@ -0,0 +1,187 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "version": "0.18.0.dev2025022500",
3
+ "pretrained_config": {
4
+ "mlp_bias": false,
5
+ "attn_bias": false,
6
+ "rotary_base": 1000000,
7
+ "rotary_scaling": {
8
+ "factor": 8.0,
9
+ "high_freq_factor": 4.0,
10
+ "low_freq_factor": 1.0,
11
+ "original_max_position_embeddings": 8192,
12
+ "rope_type": "llama3"
13
+ },
14
+ "residual_mlp": false,
15
+ "disable_weight_only_quant_plugin": false,
16
+ "moe": {
17
+ "num_experts": 0,
18
+ "shared_expert_intermediate_size": 0,
19
+ "top_k": 0,
20
+ "normalization_mode": null,
21
+ "sparse_mixer_epsilon": 0.01,
22
+ "tp_mode": 0,
23
+ "device_limited_n_group": 0,
24
+ "device_limited_topk_group": 0,
25
+ "device_limited_routed_scaling_factor": 1.0
26
+ },
27
+ "remove_duplicated_kv_heads": false,
28
+ "fc_after_embed": false,
29
+ "use_input_layernorm_in_first_layer": true,
30
+ "use_last_layernorm": true,
31
+ "layer_idx_offset": 0,
32
+ "embedding_multiplier": 1.0,
33
+ "attention_multiplier": 1.0,
34
+ "residual_multiplier": 1.0,
35
+ "output_multiplier_scale": 1.0,
36
+ "has_partial_lora_mask": false,
37
+ "architecture": "ExaoneForCausalLM",
38
+ "dtype": "bfloat16",
39
+ "vocab_size": 102400,
40
+ "hidden_size": 2560,
41
+ "num_hidden_layers": 30,
42
+ "num_attention_heads": 32,
43
+ "hidden_act": "silu",
44
+ "logits_dtype": "float32",
45
+ "norm_epsilon": 1e-05,
46
+ "runtime_defaults": null,
47
+ "position_embedding_type": "rope_gpt_neox",
48
+ "num_key_value_heads": 8,
49
+ "intermediate_size": 7168,
50
+ "max_position_embeddings": 32768,
51
+ "mapping": {
52
+ "world_size": 1,
53
+ "gpus_per_node": 8,
54
+ "cp_size": 1,
55
+ "tp_size": 1,
56
+ "pp_size": 1,
57
+ "moe_tp_size": 1,
58
+ "moe_ep_size": 1,
59
+ "auto_parallel": false
60
+ },
61
+ "quantization": {
62
+ "quant_algo": null,
63
+ "kv_cache_quant_algo": null,
64
+ "group_size": 128,
65
+ "smoothquant_val": 0.5,
66
+ "clamp_val": null,
67
+ "use_meta_recipe": false,
68
+ "has_zero_point": false,
69
+ "pre_quant_scale": false,
70
+ "exclude_modules": null
71
+ },
72
+ "use_parallel_embedding": false,
73
+ "embedding_sharding_dim": 0,
74
+ "head_size": 80,
75
+ "qk_layernorm": false,
76
+ "rotary_embedding_dim": 80,
77
+ "tie_word_embeddings": true
78
+ },
79
+ "build_config": {
80
+ "max_input_len": 1024,
81
+ "max_seq_len": 32768,
82
+ "opt_batch_size": 8,
83
+ "max_batch_size": 2048,
84
+ "max_beam_width": 1,
85
+ "max_num_tokens": 8192,
86
+ "opt_num_tokens": 2048,
87
+ "max_prompt_embedding_table_size": 0,
88
+ "kv_cache_type": "PAGED",
89
+ "gather_context_logits": false,
90
+ "gather_generation_logits": false,
91
+ "strongly_typed": true,
92
+ "force_num_profiles": null,
93
+ "profiling_verbosity": "layer_names_only",
94
+ "enable_debug_output": false,
95
+ "max_draft_len": 0,
96
+ "speculative_decoding_mode": 1,
97
+ "use_refit": false,
98
+ "input_timing_cache": null,
99
+ "output_timing_cache": "model.cache",
100
+ "lora_config": {
101
+ "lora_dir": [],
102
+ "lora_ckpt_source": "hf",
103
+ "max_lora_rank": 64,
104
+ "lora_target_modules": [],
105
+ "trtllm_modules_to_hf_modules": {}
106
+ },
107
+ "auto_parallel_config": {
108
+ "world_size": 1,
109
+ "gpus_per_node": 8,
110
+ "cluster_key": "A100-SXM-40GB",
111
+ "cluster_info": null,
112
+ "sharding_cost_model": "alpha_beta",
113
+ "comm_cost_model": "alpha_beta",
114
+ "enable_pipeline_parallelism": false,
115
+ "enable_shard_unbalanced_shape": false,
116
+ "enable_shard_dynamic_shape": false,
117
+ "enable_reduce_scatter": true,
118
+ "builder_flags": null,
119
+ "debug_mode": false,
120
+ "infer_shape": true,
121
+ "validation_mode": false,
122
+ "same_buffer_io": {
123
+ "past_key_value_(\\d+)": "present_key_value_\\1"
124
+ },
125
+ "same_spec_io": {},
126
+ "sharded_io_allowlist": [
127
+ "past_key_value_\\d+",
128
+ "present_key_value_\\d*"
129
+ ],
130
+ "fill_weights": false,
131
+ "parallel_config_cache": null,
132
+ "profile_cache": null,
133
+ "dump_path": null,
134
+ "debug_outputs": []
135
+ },
136
+ "weight_sparsity": false,
137
+ "weight_streaming": false,
138
+ "plugin_config": {
139
+ "dtype": "bfloat16",
140
+ "bert_attention_plugin": "auto",
141
+ "gpt_attention_plugin": "auto",
142
+ "gemm_plugin": "auto",
143
+ "explicitly_disable_gemm_plugin": false,
144
+ "gemm_swiglu_plugin": null,
145
+ "fp8_rowwise_gemm_plugin": null,
146
+ "qserve_gemm_plugin": null,
147
+ "identity_plugin": null,
148
+ "nccl_plugin": null,
149
+ "lora_plugin": null,
150
+ "dora_plugin": false,
151
+ "weight_only_groupwise_quant_matmul_plugin": null,
152
+ "weight_only_quant_matmul_plugin": null,
153
+ "smooth_quant_plugins": true,
154
+ "smooth_quant_gemm_plugin": null,
155
+ "layernorm_quantization_plugin": null,
156
+ "rmsnorm_quantization_plugin": null,
157
+ "quantize_per_token_plugin": false,
158
+ "quantize_tensor_plugin": false,
159
+ "moe_plugin": "auto",
160
+ "mamba_conv1d_plugin": "auto",
161
+ "low_latency_gemm_plugin": null,
162
+ "low_latency_gemm_swiglu_plugin": null,
163
+ "gemm_allreduce_plugin": null,
164
+ "context_fmha": true,
165
+ "bert_context_fmha_fp32_acc": false,
166
+ "paged_kv_cache": true,
167
+ "remove_input_padding": true,
168
+ "norm_quant_fusion": false,
169
+ "reduce_fusion": false,
170
+ "user_buffer": false,
171
+ "tokens_per_block": 32,
172
+ "use_paged_context_fmha": true,
173
+ "use_fp8_context_fmha": false,
174
+ "fuse_fp4_quant": false,
175
+ "multiple_profiles": false,
176
+ "paged_state": false,
177
+ "streamingllm": false,
178
+ "manage_weights": false,
179
+ "use_fused_mlp": true,
180
+ "pp_reduce_scatter": false
181
+ },
182
+ "use_strip_plan": false,
183
+ "max_encoder_input_len": 1024,
184
+ "monitor_memory": false,
185
+ "use_mrope": false
186
+ }
187
+ }
bf16/1-gpu/rank0.engine ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e22eae40076e0c4adf90c9ba576258d7181f3480237cc8d9e428d710b6af4a5f
3
+ size 5353570796
int4_wq/1-gpu/config.json ADDED
@@ -0,0 +1,187 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "version": "0.18.0.dev2025022500",
3
+ "pretrained_config": {
4
+ "mlp_bias": false,
5
+ "attn_bias": false,
6
+ "rotary_base": 1000000,
7
+ "rotary_scaling": {
8
+ "factor": 8.0,
9
+ "high_freq_factor": 4.0,
10
+ "low_freq_factor": 1.0,
11
+ "original_max_position_embeddings": 8192,
12
+ "rope_type": "llama3"
13
+ },
14
+ "residual_mlp": false,
15
+ "disable_weight_only_quant_plugin": false,
16
+ "moe": {
17
+ "num_experts": 0,
18
+ "shared_expert_intermediate_size": 0,
19
+ "top_k": 0,
20
+ "normalization_mode": null,
21
+ "sparse_mixer_epsilon": 0.01,
22
+ "tp_mode": 0,
23
+ "device_limited_n_group": 0,
24
+ "device_limited_topk_group": 0,
25
+ "device_limited_routed_scaling_factor": 1.0
26
+ },
27
+ "remove_duplicated_kv_heads": false,
28
+ "fc_after_embed": false,
29
+ "use_input_layernorm_in_first_layer": true,
30
+ "use_last_layernorm": true,
31
+ "layer_idx_offset": 0,
32
+ "embedding_multiplier": 1.0,
33
+ "attention_multiplier": 1.0,
34
+ "residual_multiplier": 1.0,
35
+ "output_multiplier_scale": 1.0,
36
+ "has_partial_lora_mask": false,
37
+ "architecture": "ExaoneForCausalLM",
38
+ "dtype": "float16",
39
+ "vocab_size": 102400,
40
+ "hidden_size": 2560,
41
+ "num_hidden_layers": 30,
42
+ "num_attention_heads": 32,
43
+ "hidden_act": "silu",
44
+ "logits_dtype": "float32",
45
+ "norm_epsilon": 1e-05,
46
+ "runtime_defaults": null,
47
+ "position_embedding_type": "rope_gpt_neox",
48
+ "num_key_value_heads": 8,
49
+ "intermediate_size": 7168,
50
+ "max_position_embeddings": 32768,
51
+ "mapping": {
52
+ "world_size": 1,
53
+ "gpus_per_node": 8,
54
+ "cp_size": 1,
55
+ "tp_size": 1,
56
+ "pp_size": 1,
57
+ "moe_tp_size": 1,
58
+ "moe_ep_size": 1,
59
+ "auto_parallel": false
60
+ },
61
+ "quantization": {
62
+ "quant_algo": "W4A16",
63
+ "kv_cache_quant_algo": null,
64
+ "group_size": 128,
65
+ "smoothquant_val": 0.5,
66
+ "clamp_val": null,
67
+ "use_meta_recipe": false,
68
+ "has_zero_point": false,
69
+ "pre_quant_scale": false,
70
+ "exclude_modules": null
71
+ },
72
+ "use_parallel_embedding": false,
73
+ "embedding_sharding_dim": 0,
74
+ "head_size": 80,
75
+ "qk_layernorm": false,
76
+ "rotary_embedding_dim": 80,
77
+ "tie_word_embeddings": true
78
+ },
79
+ "build_config": {
80
+ "max_input_len": 1024,
81
+ "max_seq_len": 32768,
82
+ "opt_batch_size": 8,
83
+ "max_batch_size": 2048,
84
+ "max_beam_width": 1,
85
+ "max_num_tokens": 8192,
86
+ "opt_num_tokens": 2048,
87
+ "max_prompt_embedding_table_size": 0,
88
+ "kv_cache_type": "PAGED",
89
+ "gather_context_logits": false,
90
+ "gather_generation_logits": false,
91
+ "strongly_typed": true,
92
+ "force_num_profiles": null,
93
+ "profiling_verbosity": "layer_names_only",
94
+ "enable_debug_output": false,
95
+ "max_draft_len": 0,
96
+ "speculative_decoding_mode": 1,
97
+ "use_refit": false,
98
+ "input_timing_cache": null,
99
+ "output_timing_cache": "model.cache",
100
+ "lora_config": {
101
+ "lora_dir": [],
102
+ "lora_ckpt_source": "hf",
103
+ "max_lora_rank": 64,
104
+ "lora_target_modules": [],
105
+ "trtllm_modules_to_hf_modules": {}
106
+ },
107
+ "auto_parallel_config": {
108
+ "world_size": 1,
109
+ "gpus_per_node": 8,
110
+ "cluster_key": "A100-SXM-40GB",
111
+ "cluster_info": null,
112
+ "sharding_cost_model": "alpha_beta",
113
+ "comm_cost_model": "alpha_beta",
114
+ "enable_pipeline_parallelism": false,
115
+ "enable_shard_unbalanced_shape": false,
116
+ "enable_shard_dynamic_shape": false,
117
+ "enable_reduce_scatter": true,
118
+ "builder_flags": null,
119
+ "debug_mode": false,
120
+ "infer_shape": true,
121
+ "validation_mode": false,
122
+ "same_buffer_io": {
123
+ "past_key_value_(\\d+)": "present_key_value_\\1"
124
+ },
125
+ "same_spec_io": {},
126
+ "sharded_io_allowlist": [
127
+ "past_key_value_\\d+",
128
+ "present_key_value_\\d*"
129
+ ],
130
+ "fill_weights": false,
131
+ "parallel_config_cache": null,
132
+ "profile_cache": null,
133
+ "dump_path": null,
134
+ "debug_outputs": []
135
+ },
136
+ "weight_sparsity": false,
137
+ "weight_streaming": false,
138
+ "plugin_config": {
139
+ "dtype": "float16",
140
+ "bert_attention_plugin": "auto",
141
+ "gpt_attention_plugin": "auto",
142
+ "gemm_plugin": "auto",
143
+ "explicitly_disable_gemm_plugin": false,
144
+ "gemm_swiglu_plugin": null,
145
+ "fp8_rowwise_gemm_plugin": null,
146
+ "qserve_gemm_plugin": null,
147
+ "identity_plugin": null,
148
+ "nccl_plugin": null,
149
+ "lora_plugin": null,
150
+ "dora_plugin": false,
151
+ "weight_only_groupwise_quant_matmul_plugin": null,
152
+ "weight_only_quant_matmul_plugin": "float16",
153
+ "smooth_quant_plugins": true,
154
+ "smooth_quant_gemm_plugin": null,
155
+ "layernorm_quantization_plugin": null,
156
+ "rmsnorm_quantization_plugin": null,
157
+ "quantize_per_token_plugin": false,
158
+ "quantize_tensor_plugin": false,
159
+ "moe_plugin": "auto",
160
+ "mamba_conv1d_plugin": "auto",
161
+ "low_latency_gemm_plugin": null,
162
+ "low_latency_gemm_swiglu_plugin": null,
163
+ "gemm_allreduce_plugin": null,
164
+ "context_fmha": true,
165
+ "bert_context_fmha_fp32_acc": false,
166
+ "paged_kv_cache": true,
167
+ "remove_input_padding": true,
168
+ "norm_quant_fusion": false,
169
+ "reduce_fusion": false,
170
+ "user_buffer": false,
171
+ "tokens_per_block": 32,
172
+ "use_paged_context_fmha": true,
173
+ "use_fp8_context_fmha": false,
174
+ "fuse_fp4_quant": false,
175
+ "multiple_profiles": false,
176
+ "paged_state": false,
177
+ "streamingllm": false,
178
+ "manage_weights": false,
179
+ "use_fused_mlp": true,
180
+ "pp_reduce_scatter": false
181
+ },
182
+ "use_strip_plan": false,
183
+ "max_encoder_input_len": 1024,
184
+ "monitor_memory": false,
185
+ "use_mrope": false
186
+ }
187
+ }
int4_wq/1-gpu/rank0.engine ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c16323aec56aa853bf3d6fe00bb1eda21cce4798f140b70837b441fea6589ef6
3
+ size 2141028524
int8_wq/1-gpu/config.json ADDED
@@ -0,0 +1,187 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "version": "0.18.0.dev2025022500",
3
+ "pretrained_config": {
4
+ "mlp_bias": false,
5
+ "attn_bias": false,
6
+ "rotary_base": 1000000,
7
+ "rotary_scaling": {
8
+ "factor": 8.0,
9
+ "high_freq_factor": 4.0,
10
+ "low_freq_factor": 1.0,
11
+ "original_max_position_embeddings": 8192,
12
+ "rope_type": "llama3"
13
+ },
14
+ "residual_mlp": false,
15
+ "disable_weight_only_quant_plugin": false,
16
+ "moe": {
17
+ "num_experts": 0,
18
+ "shared_expert_intermediate_size": 0,
19
+ "top_k": 0,
20
+ "normalization_mode": null,
21
+ "sparse_mixer_epsilon": 0.01,
22
+ "tp_mode": 0,
23
+ "device_limited_n_group": 0,
24
+ "device_limited_topk_group": 0,
25
+ "device_limited_routed_scaling_factor": 1.0
26
+ },
27
+ "remove_duplicated_kv_heads": false,
28
+ "fc_after_embed": false,
29
+ "use_input_layernorm_in_first_layer": true,
30
+ "use_last_layernorm": true,
31
+ "layer_idx_offset": 0,
32
+ "embedding_multiplier": 1.0,
33
+ "attention_multiplier": 1.0,
34
+ "residual_multiplier": 1.0,
35
+ "output_multiplier_scale": 1.0,
36
+ "has_partial_lora_mask": false,
37
+ "architecture": "ExaoneForCausalLM",
38
+ "dtype": "float16",
39
+ "vocab_size": 102400,
40
+ "hidden_size": 2560,
41
+ "num_hidden_layers": 30,
42
+ "num_attention_heads": 32,
43
+ "hidden_act": "silu",
44
+ "logits_dtype": "float32",
45
+ "norm_epsilon": 1e-05,
46
+ "runtime_defaults": null,
47
+ "position_embedding_type": "rope_gpt_neox",
48
+ "num_key_value_heads": 8,
49
+ "intermediate_size": 7168,
50
+ "max_position_embeddings": 32768,
51
+ "mapping": {
52
+ "world_size": 1,
53
+ "gpus_per_node": 8,
54
+ "cp_size": 1,
55
+ "tp_size": 1,
56
+ "pp_size": 1,
57
+ "moe_tp_size": 1,
58
+ "moe_ep_size": 1,
59
+ "auto_parallel": false
60
+ },
61
+ "quantization": {
62
+ "quant_algo": "W8A16",
63
+ "kv_cache_quant_algo": null,
64
+ "group_size": 128,
65
+ "smoothquant_val": 0.5,
66
+ "clamp_val": null,
67
+ "use_meta_recipe": false,
68
+ "has_zero_point": false,
69
+ "pre_quant_scale": false,
70
+ "exclude_modules": null
71
+ },
72
+ "use_parallel_embedding": false,
73
+ "embedding_sharding_dim": 0,
74
+ "head_size": 80,
75
+ "qk_layernorm": false,
76
+ "rotary_embedding_dim": 80,
77
+ "tie_word_embeddings": true
78
+ },
79
+ "build_config": {
80
+ "max_input_len": 1024,
81
+ "max_seq_len": 32768,
82
+ "opt_batch_size": 8,
83
+ "max_batch_size": 2048,
84
+ "max_beam_width": 1,
85
+ "max_num_tokens": 8192,
86
+ "opt_num_tokens": 2048,
87
+ "max_prompt_embedding_table_size": 0,
88
+ "kv_cache_type": "PAGED",
89
+ "gather_context_logits": false,
90
+ "gather_generation_logits": false,
91
+ "strongly_typed": true,
92
+ "force_num_profiles": null,
93
+ "profiling_verbosity": "layer_names_only",
94
+ "enable_debug_output": false,
95
+ "max_draft_len": 0,
96
+ "speculative_decoding_mode": 1,
97
+ "use_refit": false,
98
+ "input_timing_cache": null,
99
+ "output_timing_cache": "model.cache",
100
+ "lora_config": {
101
+ "lora_dir": [],
102
+ "lora_ckpt_source": "hf",
103
+ "max_lora_rank": 64,
104
+ "lora_target_modules": [],
105
+ "trtllm_modules_to_hf_modules": {}
106
+ },
107
+ "auto_parallel_config": {
108
+ "world_size": 1,
109
+ "gpus_per_node": 8,
110
+ "cluster_key": "A100-SXM-40GB",
111
+ "cluster_info": null,
112
+ "sharding_cost_model": "alpha_beta",
113
+ "comm_cost_model": "alpha_beta",
114
+ "enable_pipeline_parallelism": false,
115
+ "enable_shard_unbalanced_shape": false,
116
+ "enable_shard_dynamic_shape": false,
117
+ "enable_reduce_scatter": true,
118
+ "builder_flags": null,
119
+ "debug_mode": false,
120
+ "infer_shape": true,
121
+ "validation_mode": false,
122
+ "same_buffer_io": {
123
+ "past_key_value_(\\d+)": "present_key_value_\\1"
124
+ },
125
+ "same_spec_io": {},
126
+ "sharded_io_allowlist": [
127
+ "past_key_value_\\d+",
128
+ "present_key_value_\\d*"
129
+ ],
130
+ "fill_weights": false,
131
+ "parallel_config_cache": null,
132
+ "profile_cache": null,
133
+ "dump_path": null,
134
+ "debug_outputs": []
135
+ },
136
+ "weight_sparsity": false,
137
+ "weight_streaming": false,
138
+ "plugin_config": {
139
+ "dtype": "float16",
140
+ "bert_attention_plugin": "auto",
141
+ "gpt_attention_plugin": "auto",
142
+ "gemm_plugin": "auto",
143
+ "explicitly_disable_gemm_plugin": false,
144
+ "gemm_swiglu_plugin": null,
145
+ "fp8_rowwise_gemm_plugin": null,
146
+ "qserve_gemm_plugin": null,
147
+ "identity_plugin": null,
148
+ "nccl_plugin": null,
149
+ "lora_plugin": null,
150
+ "dora_plugin": false,
151
+ "weight_only_groupwise_quant_matmul_plugin": null,
152
+ "weight_only_quant_matmul_plugin": "float16",
153
+ "smooth_quant_plugins": true,
154
+ "smooth_quant_gemm_plugin": null,
155
+ "layernorm_quantization_plugin": null,
156
+ "rmsnorm_quantization_plugin": null,
157
+ "quantize_per_token_plugin": false,
158
+ "quantize_tensor_plugin": false,
159
+ "moe_plugin": "auto",
160
+ "mamba_conv1d_plugin": "auto",
161
+ "low_latency_gemm_plugin": null,
162
+ "low_latency_gemm_swiglu_plugin": null,
163
+ "gemm_allreduce_plugin": null,
164
+ "context_fmha": true,
165
+ "bert_context_fmha_fp32_acc": false,
166
+ "paged_kv_cache": true,
167
+ "remove_input_padding": true,
168
+ "norm_quant_fusion": false,
169
+ "reduce_fusion": false,
170
+ "user_buffer": false,
171
+ "tokens_per_block": 32,
172
+ "use_paged_context_fmha": true,
173
+ "use_fp8_context_fmha": false,
174
+ "fuse_fp4_quant": false,
175
+ "multiple_profiles": false,
176
+ "paged_state": false,
177
+ "streamingllm": false,
178
+ "manage_weights": false,
179
+ "use_fused_mlp": true,
180
+ "pp_reduce_scatter": false
181
+ },
182
+ "use_strip_plan": false,
183
+ "max_encoder_input_len": 1024,
184
+ "monitor_memory": false,
185
+ "use_mrope": false
186
+ }
187
+ }
int8_wq/1-gpu/rank0.engine ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c8fdd6f2c85694dddd8e01511b115265f0254fb77f55893c49679e2e9e3b3564
3
+ size 3212596484