Synchronizing local compiler cache.
Browse files- neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev3/qwen3_moe/Qwen/Qwen3-30B-A3B-Instruct-2507/fc72a21b6de27e9bcefe.json +79 -0
- neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/granite/hf-internal-testing/tiny-random-GraniteForCausalLM/6a2eef8e3936e308d719.json +70 -0
- neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/granite/hf-internal-testing/tiny-random-GraniteForCausalLM/b97e39a8f58beb7389bc.json +70 -0
- neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/granite/hf-internal-testing/tiny-random-GraniteForCausalLM/c4344d0006797e4575a8.json +70 -0
- neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/llama/llamafactory/tiny-random-Llama-3/0877b2b6ccc545a88fe8.json +74 -0
- neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/llama/llamafactory/tiny-random-Llama-3/674878857b261e357c7b.json +74 -0
- neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/llama/llamafactory/tiny-random-Llama-3/b54003e874e8671bff7d.json +74 -0
- neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/llama/unsloth/Llama-3.2-1B-Instruct/cd641ba46e0fe08cbcc1.json +75 -0
- neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/mixtral/dacorvo/Mixtral-tiny/2db16eb8799b9160edec.json +70 -0
- neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/mixtral/dacorvo/Mixtral-tiny/a87206b079f3f365cf93.json +70 -0
- neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/mixtral/dacorvo/Mixtral-tiny/ba9ad122c306f79a7f90.json +70 -0
- neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/phi3/yujiepan/phi-4-tiny-random/308e16ab3ee4911957e7.json +71 -0
- neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/phi3/yujiepan/phi-4-tiny-random/38df0e7247208af80c7d.json +71 -0
- neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/phi3/yujiepan/phi-4-tiny-random/f190d1b1acf6d9624f9c.json +71 -0
- neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/qwen2/yujiepan/qwen2.5-128k-tiny-random/68fbb28f6340b2d7ca3b.json +76 -0
- neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/qwen2/yujiepan/qwen2.5-128k-tiny-random/691934144efc9536a94c.json +76 -0
- neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/qwen2/yujiepan/qwen2.5-128k-tiny-random/945a35693afe97130cb5.json +76 -0
- neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/qwen3_moe/optimum-internal-testing/tiny-random-qwen3_moe/8ca6f1dd90ea7ad4fb33.json +77 -0
- neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/qwen3_moe/optimum-internal-testing/tiny-random-qwen3_moe/bb735c1f623df03e8d42.json +77 -0
- neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/qwen3_moe/optimum-internal-testing/tiny-random-qwen3_moe/eeeb1f6469ea444de4d6.json +77 -0
neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev3/qwen3_moe/Qwen/Qwen3-30B-A3B-Instruct-2507/fc72a21b6de27e9bcefe.json
ADDED
@@ -0,0 +1,79 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_entry_class": "SingleModelCacheEntry",
|
3 |
+
"_model_id": "Qwen/Qwen3-30B-A3B-Instruct-2507",
|
4 |
+
"_task": "text-generation",
|
5 |
+
"architectures": [
|
6 |
+
"Qwen3MoeForCausalLM"
|
7 |
+
],
|
8 |
+
"attention_bias": false,
|
9 |
+
"attention_dropout": 0.0,
|
10 |
+
"decoder_sparse_step": 1,
|
11 |
+
"head_dim": 128,
|
12 |
+
"hidden_act": "silu",
|
13 |
+
"hidden_size": 2048,
|
14 |
+
"initializer_range": 0.02,
|
15 |
+
"intermediate_size": 6144,
|
16 |
+
"max_position_embeddings": 262144,
|
17 |
+
"max_window_layers": 48,
|
18 |
+
"mlp_only_layers": [],
|
19 |
+
"model_type": "qwen3_moe",
|
20 |
+
"moe_intermediate_size": 768,
|
21 |
+
"neuron": {
|
22 |
+
"_serialized_key": "NxDNeuronConfig",
|
23 |
+
"async_mode": false,
|
24 |
+
"attn_kernel_enabled": false,
|
25 |
+
"batch_size": 8,
|
26 |
+
"capacity_factor": null,
|
27 |
+
"cc_pipeline_tiling_factor": 2,
|
28 |
+
"checkpoint_id": "Qwen/Qwen3-30B-A3B-Instruct-2507",
|
29 |
+
"checkpoint_revision": "61082d4deaa4785f64943b443cbc2b5de7524fad",
|
30 |
+
"continuous_batching": false,
|
31 |
+
"enable_bucketing": false,
|
32 |
+
"ep_degree": 1,
|
33 |
+
"flash_decoding_enabled": false,
|
34 |
+
"fused_qkv": false,
|
35 |
+
"glu_mlp": true,
|
36 |
+
"is_chunked_prefill": false,
|
37 |
+
"local_ranks_size": 8,
|
38 |
+
"logical_nc_config": 1,
|
39 |
+
"max_batch_size": 8,
|
40 |
+
"max_context_length": 4096,
|
41 |
+
"max_topk": 256,
|
42 |
+
"mlp_kernel_enabled": false,
|
43 |
+
"mlp_kernel_fuse_residual_add": false,
|
44 |
+
"n_active_tokens": 4096,
|
45 |
+
"neuronxcc_version": "2.19.8089.0+8ab9f450",
|
46 |
+
"num_cores_per_group": 1,
|
47 |
+
"on_device_sampling": false,
|
48 |
+
"optimum_neuron_version": "0.3.1.dev3",
|
49 |
+
"output_logits": false,
|
50 |
+
"padding_side": "right",
|
51 |
+
"pp_degree": 1,
|
52 |
+
"qkv_kernel_enabled": false,
|
53 |
+
"rpl_reduce_dtype": "bfloat16",
|
54 |
+
"sequence_length": 4096,
|
55 |
+
"sequence_parallel_enabled": false,
|
56 |
+
"speculation_length": 0,
|
57 |
+
"start_rank_id": 0,
|
58 |
+
"target": null,
|
59 |
+
"torch_dtype": "bfloat16",
|
60 |
+
"tp_degree": 8,
|
61 |
+
"vocab_parallel": false
|
62 |
+
},
|
63 |
+
"norm_topk_prob": true,
|
64 |
+
"num_attention_heads": 32,
|
65 |
+
"num_experts": 128,
|
66 |
+
"num_experts_per_tok": 8,
|
67 |
+
"num_hidden_layers": 48,
|
68 |
+
"num_key_value_heads": 4,
|
69 |
+
"output_router_logits": false,
|
70 |
+
"rms_norm_eps": 1e-06,
|
71 |
+
"rope_scaling": null,
|
72 |
+
"rope_theta": 10000000,
|
73 |
+
"router_aux_loss_coef": 0.001,
|
74 |
+
"sliding_window": null,
|
75 |
+
"tie_word_embeddings": false,
|
76 |
+
"use_cache": true,
|
77 |
+
"use_sliding_window": false,
|
78 |
+
"vocab_size": 151936
|
79 |
+
}
|
neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/granite/hf-internal-testing/tiny-random-GraniteForCausalLM/6a2eef8e3936e308d719.json
ADDED
@@ -0,0 +1,70 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_entry_class": "SingleModelCacheEntry",
|
3 |
+
"_model_id": "hf-internal-testing/tiny-random-GraniteForCausalLM",
|
4 |
+
"_task": "text-generation",
|
5 |
+
"architectures": [
|
6 |
+
"GraniteForCausalLM"
|
7 |
+
],
|
8 |
+
"attention_bias": false,
|
9 |
+
"attention_dropout": 0.0,
|
10 |
+
"attention_multiplier": 1.0,
|
11 |
+
"embedding_multiplier": 1.0,
|
12 |
+
"hidden_act": "silu",
|
13 |
+
"hidden_size": 32,
|
14 |
+
"initializer_range": 0.02,
|
15 |
+
"intermediate_size": 64,
|
16 |
+
"logits_scaling": 1.0,
|
17 |
+
"max_position_embeddings": 2048,
|
18 |
+
"mlp_bias": false,
|
19 |
+
"model_type": "granite",
|
20 |
+
"neuron": {
|
21 |
+
"_serialized_key": "NxDNeuronConfig",
|
22 |
+
"async_mode": false,
|
23 |
+
"attn_kernel_enabled": false,
|
24 |
+
"batch_size": 2,
|
25 |
+
"capacity_factor": null,
|
26 |
+
"cc_pipeline_tiling_factor": 2,
|
27 |
+
"checkpoint_id": "hf-internal-testing/tiny-random-GraniteForCausalLM",
|
28 |
+
"checkpoint_revision": "c3074ebc0ac2fe545305f5e5f6cce2cc9b2aa0c5",
|
29 |
+
"continuous_batching": true,
|
30 |
+
"enable_bucketing": false,
|
31 |
+
"ep_degree": 1,
|
32 |
+
"flash_decoding_enabled": false,
|
33 |
+
"fused_qkv": true,
|
34 |
+
"glu_mlp": true,
|
35 |
+
"is_chunked_prefill": false,
|
36 |
+
"local_ranks_size": 2,
|
37 |
+
"logical_nc_config": 1,
|
38 |
+
"max_batch_size": 2,
|
39 |
+
"max_context_length": 100,
|
40 |
+
"max_topk": 256,
|
41 |
+
"mlp_kernel_enabled": false,
|
42 |
+
"mlp_kernel_fuse_residual_add": false,
|
43 |
+
"n_active_tokens": 100,
|
44 |
+
"neuronxcc_version": "2.19.8089.0+8ab9f450",
|
45 |
+
"num_cores_per_group": 1,
|
46 |
+
"on_device_sampling": true,
|
47 |
+
"optimum_neuron_version": "0.3.1.dev4",
|
48 |
+
"output_logits": false,
|
49 |
+
"pp_degree": 1,
|
50 |
+
"qkv_kernel_enabled": false,
|
51 |
+
"sequence_length": 100,
|
52 |
+
"sequence_parallel_enabled": false,
|
53 |
+
"speculation_length": 0,
|
54 |
+
"start_rank_id": 0,
|
55 |
+
"target": null,
|
56 |
+
"torch_dtype": "float16",
|
57 |
+
"tp_degree": 2,
|
58 |
+
"vocab_parallel": false
|
59 |
+
},
|
60 |
+
"num_attention_heads": 4,
|
61 |
+
"num_hidden_layers": 2,
|
62 |
+
"num_key_value_heads": 4,
|
63 |
+
"residual_multiplier": 1.0,
|
64 |
+
"rms_norm_eps": 1e-06,
|
65 |
+
"rope_scaling": null,
|
66 |
+
"rope_theta": 10000.0,
|
67 |
+
"tie_word_embeddings": false,
|
68 |
+
"use_cache": true,
|
69 |
+
"vocab_size": 49152
|
70 |
+
}
|
neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/granite/hf-internal-testing/tiny-random-GraniteForCausalLM/b97e39a8f58beb7389bc.json
ADDED
@@ -0,0 +1,70 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_entry_class": "SingleModelCacheEntry",
|
3 |
+
"_model_id": "hf-internal-testing/tiny-random-GraniteForCausalLM",
|
4 |
+
"_task": "text-generation",
|
5 |
+
"architectures": [
|
6 |
+
"GraniteForCausalLM"
|
7 |
+
],
|
8 |
+
"attention_bias": false,
|
9 |
+
"attention_dropout": 0.0,
|
10 |
+
"attention_multiplier": 1.0,
|
11 |
+
"embedding_multiplier": 1.0,
|
12 |
+
"hidden_act": "silu",
|
13 |
+
"hidden_size": 32,
|
14 |
+
"initializer_range": 0.02,
|
15 |
+
"intermediate_size": 64,
|
16 |
+
"logits_scaling": 1.0,
|
17 |
+
"max_position_embeddings": 2048,
|
18 |
+
"mlp_bias": false,
|
19 |
+
"model_type": "granite",
|
20 |
+
"neuron": {
|
21 |
+
"_serialized_key": "NxDNeuronConfig",
|
22 |
+
"async_mode": false,
|
23 |
+
"attn_kernel_enabled": false,
|
24 |
+
"batch_size": 1,
|
25 |
+
"capacity_factor": null,
|
26 |
+
"cc_pipeline_tiling_factor": 2,
|
27 |
+
"checkpoint_id": "hf-internal-testing/tiny-random-GraniteForCausalLM",
|
28 |
+
"checkpoint_revision": "c3074ebc0ac2fe545305f5e5f6cce2cc9b2aa0c5",
|
29 |
+
"continuous_batching": false,
|
30 |
+
"enable_bucketing": false,
|
31 |
+
"ep_degree": 1,
|
32 |
+
"flash_decoding_enabled": false,
|
33 |
+
"fused_qkv": true,
|
34 |
+
"glu_mlp": true,
|
35 |
+
"is_chunked_prefill": false,
|
36 |
+
"local_ranks_size": 2,
|
37 |
+
"logical_nc_config": 1,
|
38 |
+
"max_batch_size": 1,
|
39 |
+
"max_context_length": 100,
|
40 |
+
"max_topk": 256,
|
41 |
+
"mlp_kernel_enabled": false,
|
42 |
+
"mlp_kernel_fuse_residual_add": false,
|
43 |
+
"n_active_tokens": 100,
|
44 |
+
"neuronxcc_version": "2.19.8089.0+8ab9f450",
|
45 |
+
"num_cores_per_group": 1,
|
46 |
+
"on_device_sampling": true,
|
47 |
+
"optimum_neuron_version": "0.3.1.dev4",
|
48 |
+
"output_logits": false,
|
49 |
+
"pp_degree": 1,
|
50 |
+
"qkv_kernel_enabled": false,
|
51 |
+
"sequence_length": 100,
|
52 |
+
"sequence_parallel_enabled": false,
|
53 |
+
"speculation_length": 0,
|
54 |
+
"start_rank_id": 0,
|
55 |
+
"target": null,
|
56 |
+
"torch_dtype": "bfloat16",
|
57 |
+
"tp_degree": 2,
|
58 |
+
"vocab_parallel": false
|
59 |
+
},
|
60 |
+
"num_attention_heads": 4,
|
61 |
+
"num_hidden_layers": 2,
|
62 |
+
"num_key_value_heads": 4,
|
63 |
+
"residual_multiplier": 1.0,
|
64 |
+
"rms_norm_eps": 1e-06,
|
65 |
+
"rope_scaling": null,
|
66 |
+
"rope_theta": 10000.0,
|
67 |
+
"tie_word_embeddings": false,
|
68 |
+
"use_cache": true,
|
69 |
+
"vocab_size": 49152
|
70 |
+
}
|
neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/granite/hf-internal-testing/tiny-random-GraniteForCausalLM/c4344d0006797e4575a8.json
ADDED
@@ -0,0 +1,70 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_entry_class": "SingleModelCacheEntry",
|
3 |
+
"_model_id": "hf-internal-testing/tiny-random-GraniteForCausalLM",
|
4 |
+
"_task": "text-generation",
|
5 |
+
"architectures": [
|
6 |
+
"GraniteForCausalLM"
|
7 |
+
],
|
8 |
+
"attention_bias": false,
|
9 |
+
"attention_dropout": 0.0,
|
10 |
+
"attention_multiplier": 1.0,
|
11 |
+
"embedding_multiplier": 1.0,
|
12 |
+
"hidden_act": "silu",
|
13 |
+
"hidden_size": 32,
|
14 |
+
"initializer_range": 0.02,
|
15 |
+
"intermediate_size": 64,
|
16 |
+
"logits_scaling": 1.0,
|
17 |
+
"max_position_embeddings": 2048,
|
18 |
+
"mlp_bias": false,
|
19 |
+
"model_type": "granite",
|
20 |
+
"neuron": {
|
21 |
+
"_serialized_key": "NxDNeuronConfig",
|
22 |
+
"async_mode": false,
|
23 |
+
"attn_kernel_enabled": false,
|
24 |
+
"batch_size": 1,
|
25 |
+
"capacity_factor": null,
|
26 |
+
"cc_pipeline_tiling_factor": 2,
|
27 |
+
"checkpoint_id": "hf-internal-testing/tiny-random-GraniteForCausalLM",
|
28 |
+
"checkpoint_revision": "c3074ebc0ac2fe545305f5e5f6cce2cc9b2aa0c5",
|
29 |
+
"continuous_batching": false,
|
30 |
+
"enable_bucketing": false,
|
31 |
+
"ep_degree": 1,
|
32 |
+
"flash_decoding_enabled": false,
|
33 |
+
"fused_qkv": true,
|
34 |
+
"glu_mlp": true,
|
35 |
+
"is_chunked_prefill": false,
|
36 |
+
"local_ranks_size": 2,
|
37 |
+
"logical_nc_config": 1,
|
38 |
+
"max_batch_size": 1,
|
39 |
+
"max_context_length": 100,
|
40 |
+
"max_topk": 256,
|
41 |
+
"mlp_kernel_enabled": false,
|
42 |
+
"mlp_kernel_fuse_residual_add": false,
|
43 |
+
"n_active_tokens": 100,
|
44 |
+
"neuronxcc_version": "2.19.8089.0+8ab9f450",
|
45 |
+
"num_cores_per_group": 1,
|
46 |
+
"on_device_sampling": true,
|
47 |
+
"optimum_neuron_version": "0.3.1.dev4",
|
48 |
+
"output_logits": false,
|
49 |
+
"pp_degree": 1,
|
50 |
+
"qkv_kernel_enabled": false,
|
51 |
+
"sequence_length": 100,
|
52 |
+
"sequence_parallel_enabled": false,
|
53 |
+
"speculation_length": 0,
|
54 |
+
"start_rank_id": 0,
|
55 |
+
"target": null,
|
56 |
+
"torch_dtype": "float16",
|
57 |
+
"tp_degree": 2,
|
58 |
+
"vocab_parallel": false
|
59 |
+
},
|
60 |
+
"num_attention_heads": 4,
|
61 |
+
"num_hidden_layers": 2,
|
62 |
+
"num_key_value_heads": 4,
|
63 |
+
"residual_multiplier": 1.0,
|
64 |
+
"rms_norm_eps": 1e-06,
|
65 |
+
"rope_scaling": null,
|
66 |
+
"rope_theta": 10000.0,
|
67 |
+
"tie_word_embeddings": false,
|
68 |
+
"use_cache": true,
|
69 |
+
"vocab_size": 49152
|
70 |
+
}
|
neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/llama/llamafactory/tiny-random-Llama-3/0877b2b6ccc545a88fe8.json
ADDED
@@ -0,0 +1,74 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_entry_class": "SingleModelCacheEntry",
|
3 |
+
"_model_id": "llamafactory/tiny-random-Llama-3",
|
4 |
+
"_task": "text-generation",
|
5 |
+
"architectures": [
|
6 |
+
"LlamaForCausalLM"
|
7 |
+
],
|
8 |
+
"attention_bias": false,
|
9 |
+
"attention_dropout": 0.0,
|
10 |
+
"head_dim": 4,
|
11 |
+
"hidden_act": "silu",
|
12 |
+
"hidden_size": 16,
|
13 |
+
"initializer_range": 0.02,
|
14 |
+
"intermediate_size": 64,
|
15 |
+
"max_position_embeddings": 131072,
|
16 |
+
"mlp_bias": false,
|
17 |
+
"model_type": "llama",
|
18 |
+
"neuron": {
|
19 |
+
"_serialized_key": "NxDNeuronConfig",
|
20 |
+
"async_mode": false,
|
21 |
+
"attn_kernel_enabled": false,
|
22 |
+
"batch_size": 1,
|
23 |
+
"capacity_factor": null,
|
24 |
+
"cc_pipeline_tiling_factor": 2,
|
25 |
+
"checkpoint_id": "llamafactory/tiny-random-Llama-3",
|
26 |
+
"checkpoint_revision": "bf2a2e3bf199ad2ee96f02a3c00246c608db22a8",
|
27 |
+
"continuous_batching": false,
|
28 |
+
"enable_bucketing": false,
|
29 |
+
"ep_degree": 1,
|
30 |
+
"flash_decoding_enabled": false,
|
31 |
+
"fused_qkv": true,
|
32 |
+
"glu_mlp": true,
|
33 |
+
"is_chunked_prefill": false,
|
34 |
+
"local_ranks_size": 2,
|
35 |
+
"logical_nc_config": 1,
|
36 |
+
"max_batch_size": 1,
|
37 |
+
"max_context_length": 100,
|
38 |
+
"max_topk": 256,
|
39 |
+
"mlp_kernel_enabled": false,
|
40 |
+
"mlp_kernel_fuse_residual_add": false,
|
41 |
+
"n_active_tokens": 100,
|
42 |
+
"neuronxcc_version": "2.19.8089.0+8ab9f450",
|
43 |
+
"num_cores_per_group": 1,
|
44 |
+
"on_device_sampling": true,
|
45 |
+
"optimum_neuron_version": "0.3.1.dev4",
|
46 |
+
"output_logits": false,
|
47 |
+
"pp_degree": 1,
|
48 |
+
"qkv_kernel_enabled": false,
|
49 |
+
"sequence_length": 100,
|
50 |
+
"sequence_parallel_enabled": false,
|
51 |
+
"speculation_length": 0,
|
52 |
+
"start_rank_id": 0,
|
53 |
+
"target": null,
|
54 |
+
"torch_dtype": "float16",
|
55 |
+
"tp_degree": 2,
|
56 |
+
"vocab_parallel": false
|
57 |
+
},
|
58 |
+
"num_attention_heads": 4,
|
59 |
+
"num_hidden_layers": 2,
|
60 |
+
"num_key_value_heads": 4,
|
61 |
+
"pretraining_tp": 1,
|
62 |
+
"rms_norm_eps": 1e-05,
|
63 |
+
"rope_scaling": {
|
64 |
+
"factor": 8.0,
|
65 |
+
"high_freq_factor": 4.0,
|
66 |
+
"low_freq_factor": 1.0,
|
67 |
+
"original_max_position_embeddings": 8192,
|
68 |
+
"rope_type": "llama3"
|
69 |
+
},
|
70 |
+
"rope_theta": 500000.0,
|
71 |
+
"tie_word_embeddings": false,
|
72 |
+
"use_cache": true,
|
73 |
+
"vocab_size": 128256
|
74 |
+
}
|
neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/llama/llamafactory/tiny-random-Llama-3/674878857b261e357c7b.json
ADDED
@@ -0,0 +1,74 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_entry_class": "SingleModelCacheEntry",
|
3 |
+
"_model_id": "llamafactory/tiny-random-Llama-3",
|
4 |
+
"_task": "text-generation",
|
5 |
+
"architectures": [
|
6 |
+
"LlamaForCausalLM"
|
7 |
+
],
|
8 |
+
"attention_bias": false,
|
9 |
+
"attention_dropout": 0.0,
|
10 |
+
"head_dim": 4,
|
11 |
+
"hidden_act": "silu",
|
12 |
+
"hidden_size": 16,
|
13 |
+
"initializer_range": 0.02,
|
14 |
+
"intermediate_size": 64,
|
15 |
+
"max_position_embeddings": 131072,
|
16 |
+
"mlp_bias": false,
|
17 |
+
"model_type": "llama",
|
18 |
+
"neuron": {
|
19 |
+
"_serialized_key": "NxDNeuronConfig",
|
20 |
+
"async_mode": false,
|
21 |
+
"attn_kernel_enabled": false,
|
22 |
+
"batch_size": 2,
|
23 |
+
"capacity_factor": null,
|
24 |
+
"cc_pipeline_tiling_factor": 2,
|
25 |
+
"checkpoint_id": "llamafactory/tiny-random-Llama-3",
|
26 |
+
"checkpoint_revision": "bf2a2e3bf199ad2ee96f02a3c00246c608db22a8",
|
27 |
+
"continuous_batching": true,
|
28 |
+
"enable_bucketing": false,
|
29 |
+
"ep_degree": 1,
|
30 |
+
"flash_decoding_enabled": false,
|
31 |
+
"fused_qkv": true,
|
32 |
+
"glu_mlp": true,
|
33 |
+
"is_chunked_prefill": false,
|
34 |
+
"local_ranks_size": 2,
|
35 |
+
"logical_nc_config": 1,
|
36 |
+
"max_batch_size": 2,
|
37 |
+
"max_context_length": 100,
|
38 |
+
"max_topk": 256,
|
39 |
+
"mlp_kernel_enabled": false,
|
40 |
+
"mlp_kernel_fuse_residual_add": false,
|
41 |
+
"n_active_tokens": 100,
|
42 |
+
"neuronxcc_version": "2.19.8089.0+8ab9f450",
|
43 |
+
"num_cores_per_group": 1,
|
44 |
+
"on_device_sampling": true,
|
45 |
+
"optimum_neuron_version": "0.3.1.dev4",
|
46 |
+
"output_logits": false,
|
47 |
+
"pp_degree": 1,
|
48 |
+
"qkv_kernel_enabled": false,
|
49 |
+
"sequence_length": 100,
|
50 |
+
"sequence_parallel_enabled": false,
|
51 |
+
"speculation_length": 0,
|
52 |
+
"start_rank_id": 0,
|
53 |
+
"target": null,
|
54 |
+
"torch_dtype": "float16",
|
55 |
+
"tp_degree": 2,
|
56 |
+
"vocab_parallel": false
|
57 |
+
},
|
58 |
+
"num_attention_heads": 4,
|
59 |
+
"num_hidden_layers": 2,
|
60 |
+
"num_key_value_heads": 4,
|
61 |
+
"pretraining_tp": 1,
|
62 |
+
"rms_norm_eps": 1e-05,
|
63 |
+
"rope_scaling": {
|
64 |
+
"factor": 8.0,
|
65 |
+
"high_freq_factor": 4.0,
|
66 |
+
"low_freq_factor": 1.0,
|
67 |
+
"original_max_position_embeddings": 8192,
|
68 |
+
"rope_type": "llama3"
|
69 |
+
},
|
70 |
+
"rope_theta": 500000.0,
|
71 |
+
"tie_word_embeddings": false,
|
72 |
+
"use_cache": true,
|
73 |
+
"vocab_size": 128256
|
74 |
+
}
|
neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/llama/llamafactory/tiny-random-Llama-3/b54003e874e8671bff7d.json
ADDED
@@ -0,0 +1,74 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_entry_class": "SingleModelCacheEntry",
|
3 |
+
"_model_id": "llamafactory/tiny-random-Llama-3",
|
4 |
+
"_task": "text-generation",
|
5 |
+
"architectures": [
|
6 |
+
"LlamaForCausalLM"
|
7 |
+
],
|
8 |
+
"attention_bias": false,
|
9 |
+
"attention_dropout": 0.0,
|
10 |
+
"head_dim": 4,
|
11 |
+
"hidden_act": "silu",
|
12 |
+
"hidden_size": 16,
|
13 |
+
"initializer_range": 0.02,
|
14 |
+
"intermediate_size": 64,
|
15 |
+
"max_position_embeddings": 131072,
|
16 |
+
"mlp_bias": false,
|
17 |
+
"model_type": "llama",
|
18 |
+
"neuron": {
|
19 |
+
"_serialized_key": "NxDNeuronConfig",
|
20 |
+
"async_mode": false,
|
21 |
+
"attn_kernel_enabled": false,
|
22 |
+
"batch_size": 1,
|
23 |
+
"capacity_factor": null,
|
24 |
+
"cc_pipeline_tiling_factor": 2,
|
25 |
+
"checkpoint_id": "llamafactory/tiny-random-Llama-3",
|
26 |
+
"checkpoint_revision": "bf2a2e3bf199ad2ee96f02a3c00246c608db22a8",
|
27 |
+
"continuous_batching": false,
|
28 |
+
"enable_bucketing": false,
|
29 |
+
"ep_degree": 1,
|
30 |
+
"flash_decoding_enabled": false,
|
31 |
+
"fused_qkv": true,
|
32 |
+
"glu_mlp": true,
|
33 |
+
"is_chunked_prefill": false,
|
34 |
+
"local_ranks_size": 2,
|
35 |
+
"logical_nc_config": 1,
|
36 |
+
"max_batch_size": 1,
|
37 |
+
"max_context_length": 100,
|
38 |
+
"max_topk": 256,
|
39 |
+
"mlp_kernel_enabled": false,
|
40 |
+
"mlp_kernel_fuse_residual_add": false,
|
41 |
+
"n_active_tokens": 100,
|
42 |
+
"neuronxcc_version": "2.19.8089.0+8ab9f450",
|
43 |
+
"num_cores_per_group": 1,
|
44 |
+
"on_device_sampling": true,
|
45 |
+
"optimum_neuron_version": "0.3.1.dev4",
|
46 |
+
"output_logits": false,
|
47 |
+
"pp_degree": 1,
|
48 |
+
"qkv_kernel_enabled": false,
|
49 |
+
"sequence_length": 100,
|
50 |
+
"sequence_parallel_enabled": false,
|
51 |
+
"speculation_length": 0,
|
52 |
+
"start_rank_id": 0,
|
53 |
+
"target": null,
|
54 |
+
"torch_dtype": "bfloat16",
|
55 |
+
"tp_degree": 2,
|
56 |
+
"vocab_parallel": false
|
57 |
+
},
|
58 |
+
"num_attention_heads": 4,
|
59 |
+
"num_hidden_layers": 2,
|
60 |
+
"num_key_value_heads": 4,
|
61 |
+
"pretraining_tp": 1,
|
62 |
+
"rms_norm_eps": 1e-05,
|
63 |
+
"rope_scaling": {
|
64 |
+
"factor": 8.0,
|
65 |
+
"high_freq_factor": 4.0,
|
66 |
+
"low_freq_factor": 1.0,
|
67 |
+
"original_max_position_embeddings": 8192,
|
68 |
+
"rope_type": "llama3"
|
69 |
+
},
|
70 |
+
"rope_theta": 500000.0,
|
71 |
+
"tie_word_embeddings": false,
|
72 |
+
"use_cache": true,
|
73 |
+
"vocab_size": 128256
|
74 |
+
}
|
neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/llama/unsloth/Llama-3.2-1B-Instruct/cd641ba46e0fe08cbcc1.json
ADDED
@@ -0,0 +1,75 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_entry_class": "SingleModelCacheEntry",
|
3 |
+
"_model_id": "unsloth/Llama-3.2-1B-Instruct",
|
4 |
+
"_task": "text-generation",
|
5 |
+
"architectures": [
|
6 |
+
"LlamaForCausalLM"
|
7 |
+
],
|
8 |
+
"attention_bias": false,
|
9 |
+
"attention_dropout": 0.0,
|
10 |
+
"head_dim": 64,
|
11 |
+
"hidden_act": "silu",
|
12 |
+
"hidden_size": 2048,
|
13 |
+
"initializer_range": 0.02,
|
14 |
+
"intermediate_size": 8192,
|
15 |
+
"max_position_embeddings": 131072,
|
16 |
+
"mlp_bias": false,
|
17 |
+
"model_type": "llama",
|
18 |
+
"neuron": {
|
19 |
+
"_serialized_key": "NxDNeuronConfig",
|
20 |
+
"async_mode": false,
|
21 |
+
"attn_kernel_enabled": false,
|
22 |
+
"batch_size": 4,
|
23 |
+
"capacity_factor": null,
|
24 |
+
"cc_pipeline_tiling_factor": 2,
|
25 |
+
"checkpoint_id": "unsloth/Llama-3.2-1B-Instruct",
|
26 |
+
"checkpoint_revision": "5a8abab4a5d6f164389b1079fb721cfab8d7126c",
|
27 |
+
"continuous_batching": true,
|
28 |
+
"enable_bucketing": false,
|
29 |
+
"ep_degree": 1,
|
30 |
+
"flash_decoding_enabled": false,
|
31 |
+
"fused_qkv": true,
|
32 |
+
"glu_mlp": true,
|
33 |
+
"is_chunked_prefill": false,
|
34 |
+
"local_ranks_size": 2,
|
35 |
+
"logical_nc_config": 1,
|
36 |
+
"max_batch_size": 4,
|
37 |
+
"max_context_length": 4096,
|
38 |
+
"max_topk": 256,
|
39 |
+
"mlp_kernel_enabled": false,
|
40 |
+
"mlp_kernel_fuse_residual_add": false,
|
41 |
+
"n_active_tokens": 4096,
|
42 |
+
"neuronxcc_version": "2.19.8089.0+8ab9f450",
|
43 |
+
"num_cores_per_group": 1,
|
44 |
+
"on_device_sampling": true,
|
45 |
+
"optimum_neuron_version": "0.3.1.dev4",
|
46 |
+
"output_logits": false,
|
47 |
+
"pp_degree": 1,
|
48 |
+
"qkv_kernel_enabled": false,
|
49 |
+
"sequence_length": 4096,
|
50 |
+
"sequence_parallel_enabled": false,
|
51 |
+
"speculation_length": 0,
|
52 |
+
"start_rank_id": 0,
|
53 |
+
"target": null,
|
54 |
+
"torch_dtype": "float16",
|
55 |
+
"tp_degree": 2,
|
56 |
+
"vocab_parallel": false
|
57 |
+
},
|
58 |
+
"num_attention_heads": 32,
|
59 |
+
"num_hidden_layers": 16,
|
60 |
+
"num_key_value_heads": 8,
|
61 |
+
"pretraining_tp": 1,
|
62 |
+
"rms_norm_eps": 1e-05,
|
63 |
+
"rope_scaling": {
|
64 |
+
"factor": 32.0,
|
65 |
+
"high_freq_factor": 4.0,
|
66 |
+
"low_freq_factor": 1.0,
|
67 |
+
"original_max_position_embeddings": 8192,
|
68 |
+
"rope_type": "llama3"
|
69 |
+
},
|
70 |
+
"rope_theta": 500000.0,
|
71 |
+
"tie_word_embeddings": true,
|
72 |
+
"unsloth_fixed": true,
|
73 |
+
"use_cache": true,
|
74 |
+
"vocab_size": 128256
|
75 |
+
}
|
neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/mixtral/dacorvo/Mixtral-tiny/2db16eb8799b9160edec.json
ADDED
@@ -0,0 +1,70 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_entry_class": "SingleModelCacheEntry",
|
3 |
+
"_model_id": "dacorvo/Mixtral-tiny",
|
4 |
+
"_task": "text-generation",
|
5 |
+
"architectures": [
|
6 |
+
"MixtralForCausalLM"
|
7 |
+
],
|
8 |
+
"attention_dropout": 0.0,
|
9 |
+
"head_dim": 32,
|
10 |
+
"hidden_act": "silu",
|
11 |
+
"hidden_size": 1024,
|
12 |
+
"initializer_range": 0.02,
|
13 |
+
"intermediate_size": 3584,
|
14 |
+
"max_position_embeddings": 1024,
|
15 |
+
"model_type": "mixtral",
|
16 |
+
"neuron": {
|
17 |
+
"_serialized_key": "NxDNeuronConfig",
|
18 |
+
"async_mode": false,
|
19 |
+
"attn_kernel_enabled": false,
|
20 |
+
"batch_size": 1,
|
21 |
+
"capacity_factor": null,
|
22 |
+
"cc_pipeline_tiling_factor": 2,
|
23 |
+
"checkpoint_id": "dacorvo/Mixtral-tiny",
|
24 |
+
"checkpoint_revision": "c557ba205ddff6ea911f4719e0d543d6c08356b6",
|
25 |
+
"continuous_batching": false,
|
26 |
+
"enable_bucketing": false,
|
27 |
+
"ep_degree": 1,
|
28 |
+
"flash_decoding_enabled": false,
|
29 |
+
"fused_qkv": false,
|
30 |
+
"glu_mlp": true,
|
31 |
+
"is_chunked_prefill": false,
|
32 |
+
"local_ranks_size": 2,
|
33 |
+
"logical_nc_config": 1,
|
34 |
+
"max_batch_size": 1,
|
35 |
+
"max_context_length": 100,
|
36 |
+
"max_topk": 256,
|
37 |
+
"mlp_kernel_enabled": false,
|
38 |
+
"mlp_kernel_fuse_residual_add": false,
|
39 |
+
"n_active_tokens": 100,
|
40 |
+
"neuronxcc_version": "2.19.8089.0+8ab9f450",
|
41 |
+
"num_cores_per_group": 1,
|
42 |
+
"on_device_sampling": false,
|
43 |
+
"optimum_neuron_version": "0.3.1.dev4",
|
44 |
+
"output_logits": false,
|
45 |
+
"pp_degree": 1,
|
46 |
+
"qkv_kernel_enabled": false,
|
47 |
+
"sequence_length": 100,
|
48 |
+
"sequence_parallel_enabled": false,
|
49 |
+
"speculation_length": 0,
|
50 |
+
"start_rank_id": 0,
|
51 |
+
"target": null,
|
52 |
+
"torch_dtype": "bfloat16",
|
53 |
+
"tp_degree": 2,
|
54 |
+
"vocab_parallel": false
|
55 |
+
},
|
56 |
+
"num_attention_heads": 32,
|
57 |
+
"num_experts_per_tok": 2,
|
58 |
+
"num_hidden_layers": 2,
|
59 |
+
"num_key_value_heads": 8,
|
60 |
+
"num_local_experts": 8,
|
61 |
+
"output_router_logits": false,
|
62 |
+
"rms_norm_eps": 1e-05,
|
63 |
+
"rope_theta": 10000.0,
|
64 |
+
"router_aux_loss_coef": 0.001,
|
65 |
+
"router_jitter_noise": 0.0,
|
66 |
+
"sliding_window": 4096,
|
67 |
+
"tie_word_embeddings": false,
|
68 |
+
"use_cache": true,
|
69 |
+
"vocab_size": 32000
|
70 |
+
}
|
neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/mixtral/dacorvo/Mixtral-tiny/a87206b079f3f365cf93.json
ADDED
@@ -0,0 +1,70 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_entry_class": "SingleModelCacheEntry",
|
3 |
+
"_model_id": "dacorvo/Mixtral-tiny",
|
4 |
+
"_task": "text-generation",
|
5 |
+
"architectures": [
|
6 |
+
"MixtralForCausalLM"
|
7 |
+
],
|
8 |
+
"attention_dropout": 0.0,
|
9 |
+
"head_dim": 32,
|
10 |
+
"hidden_act": "silu",
|
11 |
+
"hidden_size": 1024,
|
12 |
+
"initializer_range": 0.02,
|
13 |
+
"intermediate_size": 3584,
|
14 |
+
"max_position_embeddings": 1024,
|
15 |
+
"model_type": "mixtral",
|
16 |
+
"neuron": {
|
17 |
+
"_serialized_key": "NxDNeuronConfig",
|
18 |
+
"async_mode": false,
|
19 |
+
"attn_kernel_enabled": false,
|
20 |
+
"batch_size": 2,
|
21 |
+
"capacity_factor": null,
|
22 |
+
"cc_pipeline_tiling_factor": 2,
|
23 |
+
"checkpoint_id": "dacorvo/Mixtral-tiny",
|
24 |
+
"checkpoint_revision": "c557ba205ddff6ea911f4719e0d543d6c08356b6",
|
25 |
+
"continuous_batching": false,
|
26 |
+
"enable_bucketing": false,
|
27 |
+
"ep_degree": 1,
|
28 |
+
"flash_decoding_enabled": false,
|
29 |
+
"fused_qkv": false,
|
30 |
+
"glu_mlp": true,
|
31 |
+
"is_chunked_prefill": false,
|
32 |
+
"local_ranks_size": 2,
|
33 |
+
"logical_nc_config": 1,
|
34 |
+
"max_batch_size": 2,
|
35 |
+
"max_context_length": 100,
|
36 |
+
"max_topk": 256,
|
37 |
+
"mlp_kernel_enabled": false,
|
38 |
+
"mlp_kernel_fuse_residual_add": false,
|
39 |
+
"n_active_tokens": 100,
|
40 |
+
"neuronxcc_version": "2.19.8089.0+8ab9f450",
|
41 |
+
"num_cores_per_group": 1,
|
42 |
+
"on_device_sampling": false,
|
43 |
+
"optimum_neuron_version": "0.3.1.dev4",
|
44 |
+
"output_logits": false,
|
45 |
+
"pp_degree": 1,
|
46 |
+
"qkv_kernel_enabled": false,
|
47 |
+
"sequence_length": 100,
|
48 |
+
"sequence_parallel_enabled": false,
|
49 |
+
"speculation_length": 0,
|
50 |
+
"start_rank_id": 0,
|
51 |
+
"target": null,
|
52 |
+
"torch_dtype": "float16",
|
53 |
+
"tp_degree": 2,
|
54 |
+
"vocab_parallel": false
|
55 |
+
},
|
56 |
+
"num_attention_heads": 32,
|
57 |
+
"num_experts_per_tok": 2,
|
58 |
+
"num_hidden_layers": 2,
|
59 |
+
"num_key_value_heads": 8,
|
60 |
+
"num_local_experts": 8,
|
61 |
+
"output_router_logits": false,
|
62 |
+
"rms_norm_eps": 1e-05,
|
63 |
+
"rope_theta": 10000.0,
|
64 |
+
"router_aux_loss_coef": 0.001,
|
65 |
+
"router_jitter_noise": 0.0,
|
66 |
+
"sliding_window": 4096,
|
67 |
+
"tie_word_embeddings": false,
|
68 |
+
"use_cache": true,
|
69 |
+
"vocab_size": 32000
|
70 |
+
}
|
neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/mixtral/dacorvo/Mixtral-tiny/ba9ad122c306f79a7f90.json
ADDED
@@ -0,0 +1,70 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_entry_class": "SingleModelCacheEntry",
|
3 |
+
"_model_id": "dacorvo/Mixtral-tiny",
|
4 |
+
"_task": "text-generation",
|
5 |
+
"architectures": [
|
6 |
+
"MixtralForCausalLM"
|
7 |
+
],
|
8 |
+
"attention_dropout": 0.0,
|
9 |
+
"head_dim": 32,
|
10 |
+
"hidden_act": "silu",
|
11 |
+
"hidden_size": 1024,
|
12 |
+
"initializer_range": 0.02,
|
13 |
+
"intermediate_size": 3584,
|
14 |
+
"max_position_embeddings": 1024,
|
15 |
+
"model_type": "mixtral",
|
16 |
+
"neuron": {
|
17 |
+
"_serialized_key": "NxDNeuronConfig",
|
18 |
+
"async_mode": false,
|
19 |
+
"attn_kernel_enabled": false,
|
20 |
+
"batch_size": 1,
|
21 |
+
"capacity_factor": null,
|
22 |
+
"cc_pipeline_tiling_factor": 2,
|
23 |
+
"checkpoint_id": "dacorvo/Mixtral-tiny",
|
24 |
+
"checkpoint_revision": "c557ba205ddff6ea911f4719e0d543d6c08356b6",
|
25 |
+
"continuous_batching": false,
|
26 |
+
"enable_bucketing": false,
|
27 |
+
"ep_degree": 1,
|
28 |
+
"flash_decoding_enabled": false,
|
29 |
+
"fused_qkv": false,
|
30 |
+
"glu_mlp": true,
|
31 |
+
"is_chunked_prefill": false,
|
32 |
+
"local_ranks_size": 2,
|
33 |
+
"logical_nc_config": 1,
|
34 |
+
"max_batch_size": 1,
|
35 |
+
"max_context_length": 100,
|
36 |
+
"max_topk": 256,
|
37 |
+
"mlp_kernel_enabled": false,
|
38 |
+
"mlp_kernel_fuse_residual_add": false,
|
39 |
+
"n_active_tokens": 100,
|
40 |
+
"neuronxcc_version": "2.19.8089.0+8ab9f450",
|
41 |
+
"num_cores_per_group": 1,
|
42 |
+
"on_device_sampling": false,
|
43 |
+
"optimum_neuron_version": "0.3.1.dev4",
|
44 |
+
"output_logits": false,
|
45 |
+
"pp_degree": 1,
|
46 |
+
"qkv_kernel_enabled": false,
|
47 |
+
"sequence_length": 100,
|
48 |
+
"sequence_parallel_enabled": false,
|
49 |
+
"speculation_length": 0,
|
50 |
+
"start_rank_id": 0,
|
51 |
+
"target": null,
|
52 |
+
"torch_dtype": "float16",
|
53 |
+
"tp_degree": 2,
|
54 |
+
"vocab_parallel": false
|
55 |
+
},
|
56 |
+
"num_attention_heads": 32,
|
57 |
+
"num_experts_per_tok": 2,
|
58 |
+
"num_hidden_layers": 2,
|
59 |
+
"num_key_value_heads": 8,
|
60 |
+
"num_local_experts": 8,
|
61 |
+
"output_router_logits": false,
|
62 |
+
"rms_norm_eps": 1e-05,
|
63 |
+
"rope_theta": 10000.0,
|
64 |
+
"router_aux_loss_coef": 0.001,
|
65 |
+
"router_jitter_noise": 0.0,
|
66 |
+
"sliding_window": 4096,
|
67 |
+
"tie_word_embeddings": false,
|
68 |
+
"use_cache": true,
|
69 |
+
"vocab_size": 32000
|
70 |
+
}
|
neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/phi3/yujiepan/phi-4-tiny-random/308e16ab3ee4911957e7.json
ADDED
@@ -0,0 +1,71 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_entry_class": "SingleModelCacheEntry",
|
3 |
+
"_model_id": "yujiepan/phi-4-tiny-random",
|
4 |
+
"_task": "text-generation",
|
5 |
+
"architectures": [
|
6 |
+
"Phi3ForCausalLM"
|
7 |
+
],
|
8 |
+
"attention_bias": false,
|
9 |
+
"attention_dropout": 0.0,
|
10 |
+
"auto_map": {},
|
11 |
+
"embd_pdrop": 0.0,
|
12 |
+
"hidden_act": "silu",
|
13 |
+
"hidden_size": 16,
|
14 |
+
"initializer_range": 0.02,
|
15 |
+
"intermediate_size": 32,
|
16 |
+
"max_position_embeddings": 16384,
|
17 |
+
"model_type": "phi3",
|
18 |
+
"neuron": {
|
19 |
+
"_serialized_key": "NxDNeuronConfig",
|
20 |
+
"async_mode": false,
|
21 |
+
"attn_kernel_enabled": false,
|
22 |
+
"batch_size": 1,
|
23 |
+
"capacity_factor": null,
|
24 |
+
"cc_pipeline_tiling_factor": 2,
|
25 |
+
"checkpoint_id": "yujiepan/phi-4-tiny-random",
|
26 |
+
"checkpoint_revision": "18a9a1168dc97ac6d128f811925670c275610f5a",
|
27 |
+
"continuous_batching": false,
|
28 |
+
"enable_bucketing": false,
|
29 |
+
"ep_degree": 1,
|
30 |
+
"flash_decoding_enabled": false,
|
31 |
+
"fused_qkv": true,
|
32 |
+
"glu_mlp": true,
|
33 |
+
"is_chunked_prefill": false,
|
34 |
+
"local_ranks_size": 2,
|
35 |
+
"logical_nc_config": 1,
|
36 |
+
"max_batch_size": 1,
|
37 |
+
"max_context_length": 100,
|
38 |
+
"max_topk": 256,
|
39 |
+
"mlp_kernel_enabled": false,
|
40 |
+
"mlp_kernel_fuse_residual_add": false,
|
41 |
+
"n_active_tokens": 100,
|
42 |
+
"neuronxcc_version": "2.19.8089.0+8ab9f450",
|
43 |
+
"num_cores_per_group": 1,
|
44 |
+
"on_device_sampling": true,
|
45 |
+
"optimum_neuron_version": "0.3.1.dev4",
|
46 |
+
"output_logits": false,
|
47 |
+
"pp_degree": 1,
|
48 |
+
"qkv_kernel_enabled": false,
|
49 |
+
"sequence_length": 100,
|
50 |
+
"sequence_parallel_enabled": false,
|
51 |
+
"speculation_length": 0,
|
52 |
+
"start_rank_id": 0,
|
53 |
+
"target": null,
|
54 |
+
"torch_dtype": "bfloat16",
|
55 |
+
"tp_degree": 2,
|
56 |
+
"vocab_parallel": false
|
57 |
+
},
|
58 |
+
"num_attention_heads": 2,
|
59 |
+
"num_hidden_layers": 2,
|
60 |
+
"num_key_value_heads": 1,
|
61 |
+
"original_max_position_embeddings": 16384,
|
62 |
+
"partial_rotary_factor": 1.0,
|
63 |
+
"resid_pdrop": 0.0,
|
64 |
+
"rms_norm_eps": 1e-05,
|
65 |
+
"rope_scaling": null,
|
66 |
+
"rope_theta": 250000,
|
67 |
+
"sliding_window": null,
|
68 |
+
"tie_word_embeddings": false,
|
69 |
+
"use_cache": true,
|
70 |
+
"vocab_size": 100352
|
71 |
+
}
|
neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/phi3/yujiepan/phi-4-tiny-random/38df0e7247208af80c7d.json
ADDED
@@ -0,0 +1,71 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_entry_class": "SingleModelCacheEntry",
|
3 |
+
"_model_id": "yujiepan/phi-4-tiny-random",
|
4 |
+
"_task": "text-generation",
|
5 |
+
"architectures": [
|
6 |
+
"Phi3ForCausalLM"
|
7 |
+
],
|
8 |
+
"attention_bias": false,
|
9 |
+
"attention_dropout": 0.0,
|
10 |
+
"auto_map": {},
|
11 |
+
"embd_pdrop": 0.0,
|
12 |
+
"hidden_act": "silu",
|
13 |
+
"hidden_size": 16,
|
14 |
+
"initializer_range": 0.02,
|
15 |
+
"intermediate_size": 32,
|
16 |
+
"max_position_embeddings": 16384,
|
17 |
+
"model_type": "phi3",
|
18 |
+
"neuron": {
|
19 |
+
"_serialized_key": "NxDNeuronConfig",
|
20 |
+
"async_mode": false,
|
21 |
+
"attn_kernel_enabled": false,
|
22 |
+
"batch_size": 1,
|
23 |
+
"capacity_factor": null,
|
24 |
+
"cc_pipeline_tiling_factor": 2,
|
25 |
+
"checkpoint_id": "yujiepan/phi-4-tiny-random",
|
26 |
+
"checkpoint_revision": "18a9a1168dc97ac6d128f811925670c275610f5a",
|
27 |
+
"continuous_batching": false,
|
28 |
+
"enable_bucketing": false,
|
29 |
+
"ep_degree": 1,
|
30 |
+
"flash_decoding_enabled": false,
|
31 |
+
"fused_qkv": true,
|
32 |
+
"glu_mlp": true,
|
33 |
+
"is_chunked_prefill": false,
|
34 |
+
"local_ranks_size": 2,
|
35 |
+
"logical_nc_config": 1,
|
36 |
+
"max_batch_size": 1,
|
37 |
+
"max_context_length": 100,
|
38 |
+
"max_topk": 256,
|
39 |
+
"mlp_kernel_enabled": false,
|
40 |
+
"mlp_kernel_fuse_residual_add": false,
|
41 |
+
"n_active_tokens": 100,
|
42 |
+
"neuronxcc_version": "2.19.8089.0+8ab9f450",
|
43 |
+
"num_cores_per_group": 1,
|
44 |
+
"on_device_sampling": true,
|
45 |
+
"optimum_neuron_version": "0.3.1.dev4",
|
46 |
+
"output_logits": false,
|
47 |
+
"pp_degree": 1,
|
48 |
+
"qkv_kernel_enabled": false,
|
49 |
+
"sequence_length": 100,
|
50 |
+
"sequence_parallel_enabled": false,
|
51 |
+
"speculation_length": 0,
|
52 |
+
"start_rank_id": 0,
|
53 |
+
"target": null,
|
54 |
+
"torch_dtype": "float16",
|
55 |
+
"tp_degree": 2,
|
56 |
+
"vocab_parallel": false
|
57 |
+
},
|
58 |
+
"num_attention_heads": 2,
|
59 |
+
"num_hidden_layers": 2,
|
60 |
+
"num_key_value_heads": 1,
|
61 |
+
"original_max_position_embeddings": 16384,
|
62 |
+
"partial_rotary_factor": 1.0,
|
63 |
+
"resid_pdrop": 0.0,
|
64 |
+
"rms_norm_eps": 1e-05,
|
65 |
+
"rope_scaling": null,
|
66 |
+
"rope_theta": 250000,
|
67 |
+
"sliding_window": null,
|
68 |
+
"tie_word_embeddings": false,
|
69 |
+
"use_cache": true,
|
70 |
+
"vocab_size": 100352
|
71 |
+
}
|
neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/phi3/yujiepan/phi-4-tiny-random/f190d1b1acf6d9624f9c.json
ADDED
@@ -0,0 +1,71 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_entry_class": "SingleModelCacheEntry",
|
3 |
+
"_model_id": "yujiepan/phi-4-tiny-random",
|
4 |
+
"_task": "text-generation",
|
5 |
+
"architectures": [
|
6 |
+
"Phi3ForCausalLM"
|
7 |
+
],
|
8 |
+
"attention_bias": false,
|
9 |
+
"attention_dropout": 0.0,
|
10 |
+
"auto_map": {},
|
11 |
+
"embd_pdrop": 0.0,
|
12 |
+
"hidden_act": "silu",
|
13 |
+
"hidden_size": 16,
|
14 |
+
"initializer_range": 0.02,
|
15 |
+
"intermediate_size": 32,
|
16 |
+
"max_position_embeddings": 16384,
|
17 |
+
"model_type": "phi3",
|
18 |
+
"neuron": {
|
19 |
+
"_serialized_key": "NxDNeuronConfig",
|
20 |
+
"async_mode": false,
|
21 |
+
"attn_kernel_enabled": false,
|
22 |
+
"batch_size": 2,
|
23 |
+
"capacity_factor": null,
|
24 |
+
"cc_pipeline_tiling_factor": 2,
|
25 |
+
"checkpoint_id": "yujiepan/phi-4-tiny-random",
|
26 |
+
"checkpoint_revision": "18a9a1168dc97ac6d128f811925670c275610f5a",
|
27 |
+
"continuous_batching": true,
|
28 |
+
"enable_bucketing": false,
|
29 |
+
"ep_degree": 1,
|
30 |
+
"flash_decoding_enabled": false,
|
31 |
+
"fused_qkv": true,
|
32 |
+
"glu_mlp": true,
|
33 |
+
"is_chunked_prefill": false,
|
34 |
+
"local_ranks_size": 2,
|
35 |
+
"logical_nc_config": 1,
|
36 |
+
"max_batch_size": 2,
|
37 |
+
"max_context_length": 100,
|
38 |
+
"max_topk": 256,
|
39 |
+
"mlp_kernel_enabled": false,
|
40 |
+
"mlp_kernel_fuse_residual_add": false,
|
41 |
+
"n_active_tokens": 100,
|
42 |
+
"neuronxcc_version": "2.19.8089.0+8ab9f450",
|
43 |
+
"num_cores_per_group": 1,
|
44 |
+
"on_device_sampling": true,
|
45 |
+
"optimum_neuron_version": "0.3.1.dev4",
|
46 |
+
"output_logits": false,
|
47 |
+
"pp_degree": 1,
|
48 |
+
"qkv_kernel_enabled": false,
|
49 |
+
"sequence_length": 100,
|
50 |
+
"sequence_parallel_enabled": false,
|
51 |
+
"speculation_length": 0,
|
52 |
+
"start_rank_id": 0,
|
53 |
+
"target": null,
|
54 |
+
"torch_dtype": "float16",
|
55 |
+
"tp_degree": 2,
|
56 |
+
"vocab_parallel": false
|
57 |
+
},
|
58 |
+
"num_attention_heads": 2,
|
59 |
+
"num_hidden_layers": 2,
|
60 |
+
"num_key_value_heads": 1,
|
61 |
+
"original_max_position_embeddings": 16384,
|
62 |
+
"partial_rotary_factor": 1.0,
|
63 |
+
"resid_pdrop": 0.0,
|
64 |
+
"rms_norm_eps": 1e-05,
|
65 |
+
"rope_scaling": null,
|
66 |
+
"rope_theta": 250000,
|
67 |
+
"sliding_window": null,
|
68 |
+
"tie_word_embeddings": false,
|
69 |
+
"use_cache": true,
|
70 |
+
"vocab_size": 100352
|
71 |
+
}
|
neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/qwen2/yujiepan/qwen2.5-128k-tiny-random/68fbb28f6340b2d7ca3b.json
ADDED
@@ -0,0 +1,76 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_entry_class": "SingleModelCacheEntry",
|
3 |
+
"_model_id": "yujiepan/qwen2.5-128k-tiny-random",
|
4 |
+
"_task": "text-generation",
|
5 |
+
"architectures": [
|
6 |
+
"Qwen2ForCausalLM"
|
7 |
+
],
|
8 |
+
"attention_dropout": 0.0,
|
9 |
+
"hidden_act": "silu",
|
10 |
+
"hidden_size": 8,
|
11 |
+
"initializer_range": 0.02,
|
12 |
+
"intermediate_size": 16,
|
13 |
+
"layer_types": [
|
14 |
+
"full_attention",
|
15 |
+
"full_attention"
|
16 |
+
],
|
17 |
+
"max_position_embeddings": 32768,
|
18 |
+
"max_window_layers": 1,
|
19 |
+
"model_type": "qwen2",
|
20 |
+
"neuron": {
|
21 |
+
"_serialized_key": "NxDNeuronConfig",
|
22 |
+
"async_mode": false,
|
23 |
+
"attn_kernel_enabled": false,
|
24 |
+
"batch_size": 2,
|
25 |
+
"capacity_factor": null,
|
26 |
+
"cc_pipeline_tiling_factor": 2,
|
27 |
+
"checkpoint_id": "yujiepan/qwen2.5-128k-tiny-random",
|
28 |
+
"checkpoint_revision": "c8296d4ca3f87782876d2382fbb6481d1beb8ef0",
|
29 |
+
"continuous_batching": true,
|
30 |
+
"enable_bucketing": false,
|
31 |
+
"ep_degree": 1,
|
32 |
+
"flash_decoding_enabled": false,
|
33 |
+
"fused_qkv": false,
|
34 |
+
"glu_mlp": true,
|
35 |
+
"is_chunked_prefill": false,
|
36 |
+
"local_ranks_size": 2,
|
37 |
+
"logical_nc_config": 1,
|
38 |
+
"max_batch_size": 2,
|
39 |
+
"max_context_length": 100,
|
40 |
+
"max_topk": 256,
|
41 |
+
"mlp_kernel_enabled": false,
|
42 |
+
"mlp_kernel_fuse_residual_add": false,
|
43 |
+
"n_active_tokens": 100,
|
44 |
+
"neuronxcc_version": "2.19.8089.0+8ab9f450",
|
45 |
+
"num_cores_per_group": 1,
|
46 |
+
"on_device_sampling": false,
|
47 |
+
"optimum_neuron_version": "0.3.1.dev4",
|
48 |
+
"output_logits": false,
|
49 |
+
"pp_degree": 1,
|
50 |
+
"qkv_kernel_enabled": false,
|
51 |
+
"sequence_length": 100,
|
52 |
+
"sequence_parallel_enabled": false,
|
53 |
+
"speculation_length": 0,
|
54 |
+
"start_rank_id": 0,
|
55 |
+
"target": null,
|
56 |
+
"torch_dtype": "float16",
|
57 |
+
"tp_degree": 2,
|
58 |
+
"vocab_parallel": false
|
59 |
+
},
|
60 |
+
"num_attention_heads": 4,
|
61 |
+
"num_hidden_layers": 2,
|
62 |
+
"num_key_value_heads": 2,
|
63 |
+
"rms_norm_eps": 1e-06,
|
64 |
+
"rope_scaling": {
|
65 |
+
"factor": 4.0,
|
66 |
+
"original_max_position_embeddings": 32768,
|
67 |
+
"rope_type": "yarn",
|
68 |
+
"type": "yarn"
|
69 |
+
},
|
70 |
+
"rope_theta": 1000000.0,
|
71 |
+
"sliding_window": null,
|
72 |
+
"tie_word_embeddings": false,
|
73 |
+
"use_cache": true,
|
74 |
+
"use_sliding_window": false,
|
75 |
+
"vocab_size": 152064
|
76 |
+
}
|
neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/qwen2/yujiepan/qwen2.5-128k-tiny-random/691934144efc9536a94c.json
ADDED
@@ -0,0 +1,76 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_entry_class": "SingleModelCacheEntry",
|
3 |
+
"_model_id": "yujiepan/qwen2.5-128k-tiny-random",
|
4 |
+
"_task": "text-generation",
|
5 |
+
"architectures": [
|
6 |
+
"Qwen2ForCausalLM"
|
7 |
+
],
|
8 |
+
"attention_dropout": 0.0,
|
9 |
+
"hidden_act": "silu",
|
10 |
+
"hidden_size": 8,
|
11 |
+
"initializer_range": 0.02,
|
12 |
+
"intermediate_size": 16,
|
13 |
+
"layer_types": [
|
14 |
+
"full_attention",
|
15 |
+
"full_attention"
|
16 |
+
],
|
17 |
+
"max_position_embeddings": 32768,
|
18 |
+
"max_window_layers": 1,
|
19 |
+
"model_type": "qwen2",
|
20 |
+
"neuron": {
|
21 |
+
"_serialized_key": "NxDNeuronConfig",
|
22 |
+
"async_mode": false,
|
23 |
+
"attn_kernel_enabled": false,
|
24 |
+
"batch_size": 1,
|
25 |
+
"capacity_factor": null,
|
26 |
+
"cc_pipeline_tiling_factor": 2,
|
27 |
+
"checkpoint_id": "yujiepan/qwen2.5-128k-tiny-random",
|
28 |
+
"checkpoint_revision": "c8296d4ca3f87782876d2382fbb6481d1beb8ef0",
|
29 |
+
"continuous_batching": false,
|
30 |
+
"enable_bucketing": false,
|
31 |
+
"ep_degree": 1,
|
32 |
+
"flash_decoding_enabled": false,
|
33 |
+
"fused_qkv": false,
|
34 |
+
"glu_mlp": true,
|
35 |
+
"is_chunked_prefill": false,
|
36 |
+
"local_ranks_size": 2,
|
37 |
+
"logical_nc_config": 1,
|
38 |
+
"max_batch_size": 1,
|
39 |
+
"max_context_length": 100,
|
40 |
+
"max_topk": 256,
|
41 |
+
"mlp_kernel_enabled": false,
|
42 |
+
"mlp_kernel_fuse_residual_add": false,
|
43 |
+
"n_active_tokens": 100,
|
44 |
+
"neuronxcc_version": "2.19.8089.0+8ab9f450",
|
45 |
+
"num_cores_per_group": 1,
|
46 |
+
"on_device_sampling": true,
|
47 |
+
"optimum_neuron_version": "0.3.1.dev4",
|
48 |
+
"output_logits": false,
|
49 |
+
"pp_degree": 1,
|
50 |
+
"qkv_kernel_enabled": false,
|
51 |
+
"sequence_length": 100,
|
52 |
+
"sequence_parallel_enabled": false,
|
53 |
+
"speculation_length": 0,
|
54 |
+
"start_rank_id": 0,
|
55 |
+
"target": null,
|
56 |
+
"torch_dtype": "float16",
|
57 |
+
"tp_degree": 2,
|
58 |
+
"vocab_parallel": false
|
59 |
+
},
|
60 |
+
"num_attention_heads": 4,
|
61 |
+
"num_hidden_layers": 2,
|
62 |
+
"num_key_value_heads": 2,
|
63 |
+
"rms_norm_eps": 1e-06,
|
64 |
+
"rope_scaling": {
|
65 |
+
"factor": 4.0,
|
66 |
+
"original_max_position_embeddings": 32768,
|
67 |
+
"rope_type": "yarn",
|
68 |
+
"type": "yarn"
|
69 |
+
},
|
70 |
+
"rope_theta": 1000000.0,
|
71 |
+
"sliding_window": null,
|
72 |
+
"tie_word_embeddings": false,
|
73 |
+
"use_cache": true,
|
74 |
+
"use_sliding_window": false,
|
75 |
+
"vocab_size": 152064
|
76 |
+
}
|
neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/qwen2/yujiepan/qwen2.5-128k-tiny-random/945a35693afe97130cb5.json
ADDED
@@ -0,0 +1,76 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_entry_class": "SingleModelCacheEntry",
|
3 |
+
"_model_id": "yujiepan/qwen2.5-128k-tiny-random",
|
4 |
+
"_task": "text-generation",
|
5 |
+
"architectures": [
|
6 |
+
"Qwen2ForCausalLM"
|
7 |
+
],
|
8 |
+
"attention_dropout": 0.0,
|
9 |
+
"hidden_act": "silu",
|
10 |
+
"hidden_size": 8,
|
11 |
+
"initializer_range": 0.02,
|
12 |
+
"intermediate_size": 16,
|
13 |
+
"layer_types": [
|
14 |
+
"full_attention",
|
15 |
+
"full_attention"
|
16 |
+
],
|
17 |
+
"max_position_embeddings": 32768,
|
18 |
+
"max_window_layers": 1,
|
19 |
+
"model_type": "qwen2",
|
20 |
+
"neuron": {
|
21 |
+
"_serialized_key": "NxDNeuronConfig",
|
22 |
+
"async_mode": false,
|
23 |
+
"attn_kernel_enabled": false,
|
24 |
+
"batch_size": 1,
|
25 |
+
"capacity_factor": null,
|
26 |
+
"cc_pipeline_tiling_factor": 2,
|
27 |
+
"checkpoint_id": "yujiepan/qwen2.5-128k-tiny-random",
|
28 |
+
"checkpoint_revision": "c8296d4ca3f87782876d2382fbb6481d1beb8ef0",
|
29 |
+
"continuous_batching": false,
|
30 |
+
"enable_bucketing": false,
|
31 |
+
"ep_degree": 1,
|
32 |
+
"flash_decoding_enabled": false,
|
33 |
+
"fused_qkv": false,
|
34 |
+
"glu_mlp": true,
|
35 |
+
"is_chunked_prefill": false,
|
36 |
+
"local_ranks_size": 2,
|
37 |
+
"logical_nc_config": 1,
|
38 |
+
"max_batch_size": 1,
|
39 |
+
"max_context_length": 100,
|
40 |
+
"max_topk": 256,
|
41 |
+
"mlp_kernel_enabled": false,
|
42 |
+
"mlp_kernel_fuse_residual_add": false,
|
43 |
+
"n_active_tokens": 100,
|
44 |
+
"neuronxcc_version": "2.19.8089.0+8ab9f450",
|
45 |
+
"num_cores_per_group": 1,
|
46 |
+
"on_device_sampling": true,
|
47 |
+
"optimum_neuron_version": "0.3.1.dev4",
|
48 |
+
"output_logits": false,
|
49 |
+
"pp_degree": 1,
|
50 |
+
"qkv_kernel_enabled": false,
|
51 |
+
"sequence_length": 100,
|
52 |
+
"sequence_parallel_enabled": false,
|
53 |
+
"speculation_length": 0,
|
54 |
+
"start_rank_id": 0,
|
55 |
+
"target": null,
|
56 |
+
"torch_dtype": "bfloat16",
|
57 |
+
"tp_degree": 2,
|
58 |
+
"vocab_parallel": false
|
59 |
+
},
|
60 |
+
"num_attention_heads": 4,
|
61 |
+
"num_hidden_layers": 2,
|
62 |
+
"num_key_value_heads": 2,
|
63 |
+
"rms_norm_eps": 1e-06,
|
64 |
+
"rope_scaling": {
|
65 |
+
"factor": 4.0,
|
66 |
+
"original_max_position_embeddings": 32768,
|
67 |
+
"rope_type": "yarn",
|
68 |
+
"type": "yarn"
|
69 |
+
},
|
70 |
+
"rope_theta": 1000000.0,
|
71 |
+
"sliding_window": null,
|
72 |
+
"tie_word_embeddings": false,
|
73 |
+
"use_cache": true,
|
74 |
+
"use_sliding_window": false,
|
75 |
+
"vocab_size": 152064
|
76 |
+
}
|
neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/qwen3_moe/optimum-internal-testing/tiny-random-qwen3_moe/8ca6f1dd90ea7ad4fb33.json
ADDED
@@ -0,0 +1,77 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_entry_class": "SingleModelCacheEntry",
|
3 |
+
"_model_id": "optimum-internal-testing/tiny-random-qwen3_moe",
|
4 |
+
"_task": "text-generation",
|
5 |
+
"architectures": [
|
6 |
+
"Qwen3MoeForCausalLM"
|
7 |
+
],
|
8 |
+
"attention_bias": false,
|
9 |
+
"attention_dropout": 0.0,
|
10 |
+
"decoder_sparse_step": 2,
|
11 |
+
"head_dim": 32,
|
12 |
+
"hidden_act": "silu",
|
13 |
+
"hidden_size": 64,
|
14 |
+
"initializer_range": 0.02,
|
15 |
+
"intermediate_size": 128,
|
16 |
+
"max_position_embeddings": 40960,
|
17 |
+
"max_window_layers": 1,
|
18 |
+
"mlp_only_layers": [],
|
19 |
+
"model_type": "qwen3_moe",
|
20 |
+
"moe_intermediate_size": 128,
|
21 |
+
"neuron": {
|
22 |
+
"_serialized_key": "NxDNeuronConfig",
|
23 |
+
"async_mode": false,
|
24 |
+
"attn_kernel_enabled": false,
|
25 |
+
"batch_size": 2,
|
26 |
+
"capacity_factor": null,
|
27 |
+
"cc_pipeline_tiling_factor": 2,
|
28 |
+
"checkpoint_id": "optimum-internal-testing/tiny-random-qwen3_moe",
|
29 |
+
"checkpoint_revision": "e0230be2839556b44b7400a233c73c74b4abb7af",
|
30 |
+
"continuous_batching": false,
|
31 |
+
"enable_bucketing": false,
|
32 |
+
"ep_degree": 1,
|
33 |
+
"flash_decoding_enabled": false,
|
34 |
+
"fused_qkv": false,
|
35 |
+
"glu_mlp": true,
|
36 |
+
"is_chunked_prefill": false,
|
37 |
+
"local_ranks_size": 2,
|
38 |
+
"logical_nc_config": 1,
|
39 |
+
"max_batch_size": 2,
|
40 |
+
"max_context_length": 100,
|
41 |
+
"max_topk": 256,
|
42 |
+
"mlp_kernel_enabled": false,
|
43 |
+
"mlp_kernel_fuse_residual_add": false,
|
44 |
+
"n_active_tokens": 100,
|
45 |
+
"neuronxcc_version": "2.19.8089.0+8ab9f450",
|
46 |
+
"num_cores_per_group": 1,
|
47 |
+
"on_device_sampling": false,
|
48 |
+
"optimum_neuron_version": "0.3.1.dev4",
|
49 |
+
"output_logits": false,
|
50 |
+
"pp_degree": 1,
|
51 |
+
"qkv_kernel_enabled": false,
|
52 |
+
"sequence_length": 100,
|
53 |
+
"sequence_parallel_enabled": false,
|
54 |
+
"speculation_length": 0,
|
55 |
+
"start_rank_id": 0,
|
56 |
+
"target": null,
|
57 |
+
"torch_dtype": "float16",
|
58 |
+
"tp_degree": 2,
|
59 |
+
"vocab_parallel": false
|
60 |
+
},
|
61 |
+
"norm_topk_prob": true,
|
62 |
+
"num_attention_heads": 2,
|
63 |
+
"num_experts": 8,
|
64 |
+
"num_experts_per_tok": 2,
|
65 |
+
"num_hidden_layers": 2,
|
66 |
+
"num_key_value_heads": 1,
|
67 |
+
"output_router_logits": false,
|
68 |
+
"rms_norm_eps": 1e-06,
|
69 |
+
"rope_scaling": null,
|
70 |
+
"rope_theta": 1000000.0,
|
71 |
+
"router_aux_loss_coef": 0.001,
|
72 |
+
"sliding_window": null,
|
73 |
+
"tie_word_embeddings": true,
|
74 |
+
"use_cache": true,
|
75 |
+
"use_sliding_window": false,
|
76 |
+
"vocab_size": 151936
|
77 |
+
}
|
neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/qwen3_moe/optimum-internal-testing/tiny-random-qwen3_moe/bb735c1f623df03e8d42.json
ADDED
@@ -0,0 +1,77 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_entry_class": "SingleModelCacheEntry",
|
3 |
+
"_model_id": "optimum-internal-testing/tiny-random-qwen3_moe",
|
4 |
+
"_task": "text-generation",
|
5 |
+
"architectures": [
|
6 |
+
"Qwen3MoeForCausalLM"
|
7 |
+
],
|
8 |
+
"attention_bias": false,
|
9 |
+
"attention_dropout": 0.0,
|
10 |
+
"decoder_sparse_step": 2,
|
11 |
+
"head_dim": 32,
|
12 |
+
"hidden_act": "silu",
|
13 |
+
"hidden_size": 64,
|
14 |
+
"initializer_range": 0.02,
|
15 |
+
"intermediate_size": 128,
|
16 |
+
"max_position_embeddings": 40960,
|
17 |
+
"max_window_layers": 1,
|
18 |
+
"mlp_only_layers": [],
|
19 |
+
"model_type": "qwen3_moe",
|
20 |
+
"moe_intermediate_size": 128,
|
21 |
+
"neuron": {
|
22 |
+
"_serialized_key": "NxDNeuronConfig",
|
23 |
+
"async_mode": false,
|
24 |
+
"attn_kernel_enabled": false,
|
25 |
+
"batch_size": 1,
|
26 |
+
"capacity_factor": null,
|
27 |
+
"cc_pipeline_tiling_factor": 2,
|
28 |
+
"checkpoint_id": "optimum-internal-testing/tiny-random-qwen3_moe",
|
29 |
+
"checkpoint_revision": "e0230be2839556b44b7400a233c73c74b4abb7af",
|
30 |
+
"continuous_batching": false,
|
31 |
+
"enable_bucketing": false,
|
32 |
+
"ep_degree": 1,
|
33 |
+
"flash_decoding_enabled": false,
|
34 |
+
"fused_qkv": false,
|
35 |
+
"glu_mlp": true,
|
36 |
+
"is_chunked_prefill": false,
|
37 |
+
"local_ranks_size": 2,
|
38 |
+
"logical_nc_config": 1,
|
39 |
+
"max_batch_size": 1,
|
40 |
+
"max_context_length": 100,
|
41 |
+
"max_topk": 256,
|
42 |
+
"mlp_kernel_enabled": false,
|
43 |
+
"mlp_kernel_fuse_residual_add": false,
|
44 |
+
"n_active_tokens": 100,
|
45 |
+
"neuronxcc_version": "2.19.8089.0+8ab9f450",
|
46 |
+
"num_cores_per_group": 1,
|
47 |
+
"on_device_sampling": false,
|
48 |
+
"optimum_neuron_version": "0.3.1.dev4",
|
49 |
+
"output_logits": false,
|
50 |
+
"pp_degree": 1,
|
51 |
+
"qkv_kernel_enabled": false,
|
52 |
+
"sequence_length": 100,
|
53 |
+
"sequence_parallel_enabled": false,
|
54 |
+
"speculation_length": 0,
|
55 |
+
"start_rank_id": 0,
|
56 |
+
"target": null,
|
57 |
+
"torch_dtype": "float16",
|
58 |
+
"tp_degree": 2,
|
59 |
+
"vocab_parallel": false
|
60 |
+
},
|
61 |
+
"norm_topk_prob": true,
|
62 |
+
"num_attention_heads": 2,
|
63 |
+
"num_experts": 8,
|
64 |
+
"num_experts_per_tok": 2,
|
65 |
+
"num_hidden_layers": 2,
|
66 |
+
"num_key_value_heads": 1,
|
67 |
+
"output_router_logits": false,
|
68 |
+
"rms_norm_eps": 1e-06,
|
69 |
+
"rope_scaling": null,
|
70 |
+
"rope_theta": 1000000.0,
|
71 |
+
"router_aux_loss_coef": 0.001,
|
72 |
+
"sliding_window": null,
|
73 |
+
"tie_word_embeddings": true,
|
74 |
+
"use_cache": true,
|
75 |
+
"use_sliding_window": false,
|
76 |
+
"vocab_size": 151936
|
77 |
+
}
|
neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.1.dev4/qwen3_moe/optimum-internal-testing/tiny-random-qwen3_moe/eeeb1f6469ea444de4d6.json
ADDED
@@ -0,0 +1,77 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_entry_class": "SingleModelCacheEntry",
|
3 |
+
"_model_id": "optimum-internal-testing/tiny-random-qwen3_moe",
|
4 |
+
"_task": "text-generation",
|
5 |
+
"architectures": [
|
6 |
+
"Qwen3MoeForCausalLM"
|
7 |
+
],
|
8 |
+
"attention_bias": false,
|
9 |
+
"attention_dropout": 0.0,
|
10 |
+
"decoder_sparse_step": 2,
|
11 |
+
"head_dim": 32,
|
12 |
+
"hidden_act": "silu",
|
13 |
+
"hidden_size": 64,
|
14 |
+
"initializer_range": 0.02,
|
15 |
+
"intermediate_size": 128,
|
16 |
+
"max_position_embeddings": 40960,
|
17 |
+
"max_window_layers": 1,
|
18 |
+
"mlp_only_layers": [],
|
19 |
+
"model_type": "qwen3_moe",
|
20 |
+
"moe_intermediate_size": 128,
|
21 |
+
"neuron": {
|
22 |
+
"_serialized_key": "NxDNeuronConfig",
|
23 |
+
"async_mode": false,
|
24 |
+
"attn_kernel_enabled": false,
|
25 |
+
"batch_size": 1,
|
26 |
+
"capacity_factor": null,
|
27 |
+
"cc_pipeline_tiling_factor": 2,
|
28 |
+
"checkpoint_id": "optimum-internal-testing/tiny-random-qwen3_moe",
|
29 |
+
"checkpoint_revision": "e0230be2839556b44b7400a233c73c74b4abb7af",
|
30 |
+
"continuous_batching": false,
|
31 |
+
"enable_bucketing": false,
|
32 |
+
"ep_degree": 1,
|
33 |
+
"flash_decoding_enabled": false,
|
34 |
+
"fused_qkv": false,
|
35 |
+
"glu_mlp": true,
|
36 |
+
"is_chunked_prefill": false,
|
37 |
+
"local_ranks_size": 2,
|
38 |
+
"logical_nc_config": 1,
|
39 |
+
"max_batch_size": 1,
|
40 |
+
"max_context_length": 100,
|
41 |
+
"max_topk": 256,
|
42 |
+
"mlp_kernel_enabled": false,
|
43 |
+
"mlp_kernel_fuse_residual_add": false,
|
44 |
+
"n_active_tokens": 100,
|
45 |
+
"neuronxcc_version": "2.19.8089.0+8ab9f450",
|
46 |
+
"num_cores_per_group": 1,
|
47 |
+
"on_device_sampling": false,
|
48 |
+
"optimum_neuron_version": "0.3.1.dev4",
|
49 |
+
"output_logits": false,
|
50 |
+
"pp_degree": 1,
|
51 |
+
"qkv_kernel_enabled": false,
|
52 |
+
"sequence_length": 100,
|
53 |
+
"sequence_parallel_enabled": false,
|
54 |
+
"speculation_length": 0,
|
55 |
+
"start_rank_id": 0,
|
56 |
+
"target": null,
|
57 |
+
"torch_dtype": "bfloat16",
|
58 |
+
"tp_degree": 2,
|
59 |
+
"vocab_parallel": false
|
60 |
+
},
|
61 |
+
"norm_topk_prob": true,
|
62 |
+
"num_attention_heads": 2,
|
63 |
+
"num_experts": 8,
|
64 |
+
"num_experts_per_tok": 2,
|
65 |
+
"num_hidden_layers": 2,
|
66 |
+
"num_key_value_heads": 1,
|
67 |
+
"output_router_logits": false,
|
68 |
+
"rms_norm_eps": 1e-06,
|
69 |
+
"rope_scaling": null,
|
70 |
+
"rope_theta": 1000000.0,
|
71 |
+
"router_aux_loss_coef": 0.001,
|
72 |
+
"sliding_window": null,
|
73 |
+
"tie_word_embeddings": true,
|
74 |
+
"use_cache": true,
|
75 |
+
"use_sliding_window": false,
|
76 |
+
"vocab_size": 151936
|
77 |
+
}
|